JP3567260B2 - Image data matching apparatus, image data matching method, and storage medium storing image data matching processing program - Google Patents

Description

【００６４】
そして、算出された全てのΔｉが所定値内に収まるか否かを判定し、全てが収まれば登録指紋Ａと照合指紋Ｂとは一致すると判定し、一方、そうでない場合には登録指紋Ａと照合指紋Ｂとは一致しないと判定する（ステップＳ７）。なお、ここで用いられる所定値は、例えば、複数人より取得した指紋画像データよりΔｉを実際に計算し、その計算結果の分布に基づいて所望の照合精度が得られる値とする。
【００６５】
ここで、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ７→終わり）。
【００６６】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、続いて、テンプレート配置処理部２５において、前記照合指紋回転画像メモリ１３ｃに記憶されているθｍ回転後の照合指紋画像Ｂに対し、図８（ｂ）に示すように、複数の矩形テンプレート領域ｂ_ｉ［ｉ＝１〜Ｎ：Ｎ≧２の整数（この場合Ｎ＝５）］が定義されて配置され、図９（Ａ）に示すように、その各テンプレート領域ｂｉの位置を示す座標データが照合指紋位置メモリ１３ｅに記憶される（ステップＳ７→Ｓ８）。
【００６７】
すると、照合処理部２３内の最大相関領域検出処理部２６において、前記θｍ回転後の照合指紋画像Ｂに対し定義配置された各矩形テンプレート領域ｂｉに相当する部分画像が、前記登録指紋データ記憶部２２において記憶装置１２に記憶登録された登録指紋画像Ａ上で水平方向及び垂直方向に１画素ずつラスター走査されながら、順次該テンプレートｂｉ内の全ての画素データとそれに対応する登録指紋画像Ａ内の画素データとを用いた相関係数が計算される。そして、その相関係数が最大となる登録指紋画像Ａ内の領域ａｉが、図８（ａ）に示すように検出され、図９（Ｂ）に示すように、その各最大相関領域ａｉの検出位置を示す座標データが登録指紋位置メモリ１３ｄに記憶される（ステップＳ９，Ｓ１０）。なお、相関係数の算出については後述する。
【００６８】
このθｍ回転後の照合指紋画像Ｂ上に定義した各矩形テンプレート領域ｂｉの画像データを基準とする登録指紋画像Ａ上での最大相関係数の算出及びその領域ａｉの検出は、各矩形テンプレート領域ｂｉ毎に順次行われ（ステップＳ９〜Ｓ１１）、全ての矩形テンプレート領域ｂｉにそれぞれ最大の相関係数を有する登録指紋画像Ａ上の各領域ａｉが検出されたと判断されると、照合処理部２３内の照合判定処理部２７において、図８に示すように、回転後の照合指紋画像Ｂ内に定義した前記複数の矩形テンプレート領域｛ｂ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）と登録指紋画像Ａ上から検出した前記複数の領域｛ａ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）とが比較され、該登録指紋Ａと照合指紋Ｂの同一性が評価される（ステップＳ１１→Ｓ１２）。
【００６９】
すなわち、この照合判定処理（ステップＳ１２）では、θｍ回転後照合指紋画像Ｂに配置された矩形テンプレート領域ｂ１と登録指紋画像Ａ上で検出された最大相関矩形領域ａ１の各画像上の位置を特定する座標ｂ１（Ｘ０，Ｙ０）、ａ１（ＸＤ０，ＹＤ０）と、θｍ回転後照合指紋画像Ｂに配置された各矩形テンプレート領域ｂ２〜ｂ５と登録指紋画像Ａ上で検出された各最大相関矩形領域ａ２〜ａ５それぞれの各画像上の位置を特定する座標ｂ２（Ｘ１，Ｙ１）〜ｂ５（Ｘ４，Ｙ４）、ａ２（ＸＤ１，ＹＤ１）〜ａ５（ＸＤ４，ＹＤ４）を用い、ｂ１とｂ２，ｂ３，ｂ４，ｂ５との各相対距離とａ１とａ２，ａ３，ａ４，ａ５との各相対距離との違いであるΔｉ（ｉ＝１，２，３，４）を、前記式に基づいて算出する。
【００７０】
そして、算出された全てのΔｉが所定値内に収まるか否かを判定し、全てが収まれば登録指紋Ａと照合指紋Ｂとは一致すると判定し、一方、そうでない場合には登録指紋Ａと照合指紋Ｂとは一致しないと判定する（ステップＳ１３）。
【００７１】
ここで、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ１３→終わり）。
【００７２】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、前記回転変換処理部２４において規定した全ての回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）により回転変換された各回転後照合指紋画像Ｂについて、前記ステップＳ１〜Ｓ１３に係る登録指紋画像Ａとの照合処理が実施されたか判断される（ステップＳ１３→Ｓ１４）。
【００７３】
そして、前記規定した全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについて、登録指紋画像Ａとの一連の照合処理が実施されてないと判断された場合には、前記ステップＳ１からの処理に戻り、照合指紋画像を次の回転ずれ補正角度θｍ（＝ｍ＋１）により回転変換処理し、回転された照合指紋画像Ｂについて、前記同様に、登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置した場合の照合処理、当該回転後の照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置した場合の照合処理が繰り返し実施される（ステップＳ１４→Ｓ１〜Ｓ１３）。
【００７４】
また、全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについて登録指紋画像Ａとの一連の照合処理が実施されたと判断された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋でないことが表示される（ステップＳ１４→終わり）。
【００７５】
次に、前記図４，図５に示した指紋照合処理のステップＳ３〜Ｓ５及びステップＳ９〜Ｓ１１で用いられる相関係数の算出について説明する。なお、ここでは、矩形領域Ａと矩形領域Ｂとの間の相関係数の算出として説明する。
【００７６】
まず、矩形領域Ａと矩形領域Ｂのそれぞれに含まれる画素をそれぞれ、Ａ（ｉ，ｊ）、Ｂ（ｍ，ｎ）とする。但し、矩形領域Ａと矩形領域Ｂのそれぞれに含まれる画素の総数は等しくする。また、これらの画素についての濃淡を示す多階調値である信号強度をそれぞれＸ_{ｉｊ ，}Ｙ_ｍｎとする。
【００７７】
これらの信号強度を一般化してＺｐｑと表したとき、以下の式を定義する。
【００７８】
【数２】

【００７９】
上式において、Ｎはその矩形領域に含まれる画素の総数を示す。また、上式において、Σはその矩形領域に含まれる画素の全てについての総和であることを示す。つまり、上式はその矩形領域に含まれる画素についての信号強度の平均値を示すものである。
【００８０】
次に、以下の式を更に定義する。
【００８１】
【数３】

【００８２】
上式はその矩形領域に含まれる画素についての信号強度の２乗平均値を示すものである。
【００８３】
ここで、矩形領域Ａと矩形領域Ｂとの間の相関係数Ｃ_ＡＢは、前述の式の定義を用いて表される次式により算出できる。
【００８４】
【数４】

【００８５】
上式を用いて領域間の相関係数を算出する。
【００８６】
したがって、前記構成の第１実施形態の指紋データ照合装置によれば、指紋画像を採取する際に生じた回転ずれの回転中心と回転ずれを考慮して回転ずれ補正を実施する際の回転変換処理における規定の回転中心とが一致しない場合が多いこと、及びこれにより登録指紋画像Ａ上にテンプレートを配置して照合処理を実施した場合に照合一致が得られた回転ずれ補正角度と、照合指紋画像Ｂ上にテンプレートを配置して照合処理を実施した場合に照合一致が得られた回転ずれ補正角度との絶対値が等しくならないことに着目し、回転ずれ補正角度θ１＝０°も含め、各回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）によって回転変換した照合指紋画像Ｂについて、登録指紋画像Ａ内にテンプレート領域Ａｉを配置した場合の照合処理と、回転変換した照合指紋画像Ｂ内にテンプレート領域ｂｉを配置した場合の照合処理との２通りの照合処理を交互に実施するようにしたので、回転ずれを有する同一人物・同一指の指紋照合を行うのに、より小さな回転ずれ補正角度θｍにおいて照合一致が得られる可能性が高くなり、その演算処理量を少なくして照合時間を短縮できるようになる。
【００８７】
なお、前記第１実施形態では、照合指紋画像Ｂに対してのみ回転ずれ補正のための回転変換処理を施し、回転ずれ補正角度θｍによって順次回転変換される照合指紋画像Ｂについて、登録指紋画像Ａ上にテンプレートＡｉを定義配置した場合と、当該回転変換された照合指紋画像Ｂ上にテンプレートｂｉを定義配置した場合との、２通りの照合処理を交互に実施する構成としたが、次の第２実施形態において説明するように、回転ずれ補正角度θｍで順次回転変換される照合指紋画像Ｂについて非回転の登録指紋画像Ａ上にテンプレートＡｉを定義配置して行う照合処理と、回転ずれ補正角度θｍで順次回転変換される登録指紋画像Ａについて非回転の照合指紋画像Ｂ上にテンプレートｂｉを定義配置して行う照合処理とを交互に実施する構成とし、前記同様に、より小さな回転ずれ補正角度θｍにおいて照合一致が得られるようにしてもよい。
【００８８】
（第２実施形態）
図１０は前記指紋データ照合装置の第２実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図である。
【００８９】
この第２実施形態の指紋照合動作機能では、回転変換処理部２４において、指紋データ入力部２１から与えられる照合指紋画像Ｂと、登録指紋データ記憶部２２から与えられる登録指紋画像Ａとを、回転ずれ補正角度θｍによって交互に回転変換するのと共に、照合指紋画像Ｂが回転変換されるときには、テンプレート配置処理部２５において、登録指紋データ記憶部２２から与えられる登録指紋画像Ａに対してテンプレート領域Ａｉを定義配置し、登録指紋画像Ａが回転変換されるときには、指紋データ入力部２１から与えられる照合指紋画像Ｂに対してテンプレート領域ｂｉを定義配置する。
【００９０】
そして、最大相関領域検出処理部２６及び照合判定処理部２７では、前記回転変換処理部２４において順次回転変換される照合指紋画像Ｂに対して、テンプレート配置処理部２５において登録指紋画像Ａ上に定義配置されたテンプレート領域Ａｉに対応する最大相関領域Ｂｉの検出及びその照合を行う処理と、前記回転変換処理部２４において順次回転変換される登録指紋画像Ａに対して、テンプレート配置処理部２５において照合指紋画像Ｂ上に定義配置されたテンプレート領域ｂｉに対応する最大相関領域ａｉの検出及びその照合を行う処理とを、交互に実施する構成とする。
【００９１】
次に、前記構成による指紋データ照合装置における第２実施形態の指紋照合機能について説明する。
【００９２】
図１１は前記指紋データ照合装置の第２実施形態の指紋照合処理（その１）を示すフローチャートである。
【００９３】
図１２は前記指紋データ照合装置の第２実施形態の指紋照合処理（その２）を示すフローチャートである。
【００９４】
まず、指紋データ入力部２１によって採取記憶された多階調画像である照合指紋画像Ｂの全体が、回転変換処理部２４において回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）だけ回転変換される（ステップＳ１）。但し、θ１＝０の時は、実際の回転変換は実施されない。
【００９５】
すると、テンプレート配置処理部２５において、登録指紋データ記憶部２２にて記憶された非回転の登録指紋画像Ａに対し、複数の矩形テンプレート領域Ａ_ｉ［ｉ＝１〜Ｎ：Ｎ≧２の整数（この場合Ｎ＝５）］が定義されて配置される（ステップＳ２）。
【００９６】
すると、照合処理部２３内の最大相関領域検出処理部２６において、前記登録指紋画像Ａに対し定義配置された各矩形テンプレート領域Ａｉに相当する部分画像が、前記回転変換処理部２４においてθｍ回転変換された照合指紋画像Ｂ上で水平方向及び垂直方向に１画素ずつラスター走査されながら、順次該テンプレートＡｉ内の全ての画素データとそれに対応する照合指紋画像Ｂ内の画素データとを用いた相関係数が計算される。そして、その相関係数が最大となる照合指紋画像Ｂ内の領域Ｂｉが検出される（ステップＳ３，Ｓ４）。
【００９７】
この非回転の登録指紋画像Ａ上に定義した各矩形テンプレート領域Ａｉの画像データを基準とする回転後照合指紋画像Ｂ上での最大相関係数の算出及びその領域Ｂｉの検出は、各矩形テンプレート領域Ａｉ毎に順次行われ（ステップＳ３〜Ｓ５）、全ての矩形テンプレート領域Ａｉにそれぞれ最大の相関係数を有する照合指紋画像Ｂ上の各領域Ｂｉが検出されたと判断されると、照合処理部２３内の照合判定処理部２７において、登録指紋画像Ａ内に定義した前記複数の矩形テンプレート領域｛Ａ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）と照合指紋画像Ｂ上から検出した前記複数の領域｛Ｂ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）とが比較され、該登録指紋Ａと照合指紋Ｂの同一性が評価される（ステップＳ５→Ｓ６）。
【００９８】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ７→終わり）。
【００９９】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、続いて、登録指紋データ記憶部２２に記憶されている登録指紋画像Ａの全体が、回転変換処理部２４において回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）だけ回転変換される（ステップＳ７→Ｓ８ａ）。但し、θ１＝０の時は、実際の回転変換は実施されない。
【０１００】
すると、テンプレート配置処理部２５において、前記指紋データ入力部２１により採取記憶された非回転の照合指紋画像Ｂに対し、複数の矩形テンプレート領域ｂ_ｉ［ｉ＝１〜Ｎ：Ｎ≧２の整数（この場合Ｎ＝５）］が定義されて配置される（ステップＳ８）。
【０１０１】
すると、照合処理部２３内の最大相関領域検出処理部２６において、前記非回転の照合指紋画像Ｂに対し定義配置された各矩形テンプレート領域ｂｉに相当する部分画像が、前記回転変換処理部２４によりθｍ回転変換された登録指紋画像Ａ上で水平方向及び垂直方向に１画素ずつラスター走査されながら、順次該テンプレートｂｉ内の全ての画素データとそれに対応する登録指紋画像Ａ内の画素データとを用いた相関係数が計算される。そして、その相関係数が最大となる登録指紋画像Ａ内の領域ａｉが検出される（ステップＳ９，Ｓ１０）。
【０１０２】
この非回転の照合指紋画像Ｂ上に定義した各矩形テンプレート領域ｂｉの画像データを基準とする回転後登録指紋画像Ａ上での最大相関係数の算出及びその領域ａｉの検出は、各矩形テンプレート領域ｂｉ毎に順次行われ（ステップＳ９〜Ｓ１１）、全ての矩形テンプレート領域ｂｉにそれぞれ最大の相関係数を有する登録指紋画像Ａ上の各領域ａｉが検出されたと判断されると、照合処理部２３内の照合判定処理部２７において、非回転の照合指紋画像Ｂ内に定義した前記複数の矩形テンプレート領域｛ｂ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）と回転後の登録指紋画像Ａ上から検出した前記複数の領域｛ａ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）とが比較され、該登録指紋Ａと照合指紋Ｂの同一性が評価される（ステップＳ１１→Ｓ１２）。
【０１０３】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ１３→終わり）。
【０１０４】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、前記回転変換処理部２４において規定した全ての回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）により回転変換された各回転後照合指紋画像Ｂ及び各回転後登録指紋画像Ａのそれぞれについて、前記ステップＳ１〜Ｓ１３に係る非回転の登録指紋画像Ａとの照合処理及び非回転の照合指紋画像Ｂとの照合処理が実施されたか判断される（ステップＳ１３→Ｓ１４）。
【０１０５】
そして、前記規定した全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについての非回転の登録指紋画像Ａとの照合処理及び同各回転後の登録指紋画像Ａについての非回転の照合指紋画像Ｂとの照合処理が実施されてないと判断された場合には、前記ステップＳ１からの処理に戻り、次の回転ずれ補正角度θｍ（＝ｍ＋１）により回転変換された照合指紋画像Ｂについての非回転の登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置した場合の照合処理、及び次の回転ずれ補正角度θｍ（＝ｍ＋１）により回転変換された登録指紋画像Ａについての非回転の照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置した場合の照合処理が繰り返し実施される（ステップＳ１４→Ｓ１〜Ｓ１３）。
【０１０６】
また、全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについての非回転の登録指紋画像Ａとの照合処理及び同各回転後の登録指紋画像Ａについての非回転の照合指紋画像Ｂとの照合処理が実施されたと判断された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋でないことが表示される（ステップＳ１４→終わり）。
【０１０７】
したがって、この第２実施形態の指紋データ照合装置によっても、回転ずれを有する同一人物・同一指の指紋照合を行うのに、より小さな回転ずれ補正角度θｍにおいて照合一致が得られる可能性が高くなり、その演算処理量を少なくして照合時間を短縮できるようになる。
【０１０８】
なお、前記第２実施形態では、照合指紋画像Ｂを回転変換したときには非回転の登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置して当該回転後の照合指紋画像Ｂとの照合処理を行うのと共に、登録指紋画像Ａを回転変換したときには非回転の照合指紋画像Ｂ上にテンプレートｂｉを定義配置して当該回転後の登録指紋画像Ａとの照合処理を行う構成としたが、次の第３実施形態において説明するように、照合指紋画像Ｂと登録指紋画像Ａとを交互に回転変換すると共に、その回転変換した照合指紋画像Ｂ及び登録指紋画像Ａに対してテンプレート領域ｂｉ，Ａｉを定義配置し、非回転の登録指紋画像Ａとの照合処理及び非回転の照合指紋画像Ｂとの照合処理を交互に行う構成とし、前記同様に、より小さな回転ずれ補正角度θｍにおいて照合一致が得られるようにしてもよい。
【０１０９】
（第３実施形態）
図１３は前記指紋データ照合装置の第３実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図である。
【０１１０】
この第３実施形態の指紋照合動作機能では、回転変換処理部２４において、指紋データ入力部２１から与えられる照合指紋画像Ｂと、登録指紋データ記憶部２２から与えられる登録指紋画像Ａとを、回転ずれ補正角度θｍによって交互に回転変換するのに伴い、その回転変換された照合指紋画像Ｂ及び登録指紋画像Ａに対してテンプレート領域ｂｉ及びＡｉを定義配置する。
【０１１１】
そして、最大相関領域検出処理部２６及び照合判定処理部２７では、登録指紋データ記憶部２２から与えられる非回転の登録指紋画像Ａに対して、テンプレート配置処理部２５において各回転後の照合指紋画像Ｂ上に定義配置されたテンプレート領域ｂｉに対応する最大相関領域ａｉの検出及びその照合を行う処理と、指紋データ入力部２１から与えられる非回転の照合指紋画像Ｂに対して、テンプレート配置処理部２５において各回転後の登録指紋画像Ａ上に定義配置されたテンプレート領域Ａｉに対応する最大相関領域Ｂｉの検出及びその照合を行う処理とを、交互に実施する構成とする。
【０１１２】
したがって、この第３実施形態の指紋データ照合装置によっても、回転ずれを有する同一人物・同一指の指紋照合を行うのに、より小さな回転ずれ補正角度θｍにおいて照合一致が得られる可能性が高くなり、その演算処理量を少なくして照合時間を短縮できるようになる。
【０１１３】
また、前記第２実施形態では、照合指紋画像Ｂを回転変換したときに非回転の登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置して該回転後の照合指紋画像Ｂとの照合を行う処理と、登録指紋画像Ａを回転変換したときに非回転の照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置して該回転後の登録指紋画像Ａとの照合を行う処理とを、指紋照合の一致判断が得られない状態では、その回転ずれ補正角度θｍを順次更新させて交互に繰り返し行う構成としたが、次の第４実施形態において説明するように、初期設定の回転ずれ補正角度θｍ（θ１＝０°）のときに、登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置した場合の照合指紋画像Ｂに対する照合処理にて計算される最大相関係数ＭＡＸＣＣと、照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置した場合の登録指紋画像Ａに対する照合処理にて計算される最大相関係数ｍａｘｃｃとを大小比較する。そして、登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置した場合の最大相関係数ＭＡＸＣＣの方が大きい場合には、照合指紋画像Ｂを回転ずれ補正角度θｍによって順次回転変換しながら登録指紋画像Ａ上にテンプレート領域Ａｉを定義して行う照合処理を選択して指紋照合を実施し、照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置した場合の最大相関係数ｍａｘｃｃの方が大きい場合には、登録指紋画像Ａを回転ずれ補正角度θｍによって順次回転変換しながら照合指紋画像Ｂ上にテンプレート領域ｂｉを定義して行う照合処理を選択して指紋照合を実施する構成とすることで、さらに少ない演算処理量で照合時間を短縮するようにしてもよい。
【０１１４】
（第４実施形態）
図１４は前記指紋データ照合装置の第４実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図である。
【０１１５】
この第４実施形態の指紋照合動作機能では、指紋データ入力部２１から与えられる照合指紋画像Ｂが回転変換処理部２４において回転ずれ補正角度θｍで回転変換されるときには、登録指紋データ記憶部２２から与えられる登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置し、最大相関領域検出処理部２６及び照合判定処理部２７にて前記回転された照合指紋画像Ｂ上で最大相関係数が算出される領域を検出して指紋照合を行い、また、登録指紋データ記憶部２２から与えられる登録指紋画像Ａが回転変換処理部２４において回転ずれ補正角度θｍで回転変換されるときには、指紋データ入力部２１から与えられる照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置し、最大相関領域検出処理部２６及び照合判定処理部２７にて前記回転された登録指紋画像Ａ上で最大相関係数が算出される領域を検出して指紋照合を行うものとする。この場合、前記回転変換処理部２４における回転ずれ補正角度θｍが初期設定に伴うθ１＝０°のときに、前記最大相関領域検出処理部２６において算出される登録指紋画像Ａ上にテンプレート領域Ａ１を配置した場合の照合指紋画像Ｂに対する最大相関係数ＭＡＸＣＣと、照合指紋画像Ｂ上にテンプレート領域ｂ１を配置した場合の登録指紋画像Ａに対する最大相関係数ｍａｘｃｃとを、テンプレート配置指紋データ選択処理部２９において大小比較判断し、テンプレート配置指紋データ選択処理部２９では、登録指紋画像Ａ上にテンプレート領域Ａ１を定義配置した場合の最大相関係数ＭＡＸＣＣの方が大きいと判断したときには、回転ずれ補正角度θｍを更新させたθ２＝１°以降の処理において、指紋データ入力部２１から得られる照合指紋画像Ｂを回転変換処理部２４に固定して与えるのと共に、登録指紋データ記憶部２２から得られる登録指紋画像Ａをテンプレート配置処理部２５に固定して与えるよう選択決定し、登録指紋画像Ａにテンプレート領域Ａｉを定義配置して順次回転変換される照合指紋画像Ｂとの照合を行う処理に固定する。また、照合指紋画像Ｂ上にテンプレート領域ｂ１を定義配置した場合の最大相関係数ｍａｘｃｃの方が大きいと判断したときには、回転ずれ補正角度θｍを更新させたθ２＝１°以降の処理において、登録指紋データ記憶部２２から得られる登録指紋画像Ａを回転変換処理部２４に固定して与えるのと共に、指紋データ入力部２１から得られる照合指紋画像Ｂをテンプレート配置処理部２５に固定して与えるよう選択決定し、照合指紋画像Ｂにテンプレート領域ｂｉを定義配置して順次回転変換される登録指紋画像Ａとの照合を行う処理に固定する。
【０１１６】
次に、前記構成による指紋データ照合装置における第４実施形態の指紋照合機能について説明する。
【０１１７】
図１５は前記指紋データ照合装置の第４実施形態の指紋照合処理（その１）を示すフローチャートである。
【０１１８】
図１６は前記指紋データ照合装置の第４実施形態の指紋照合処理（その２）を示すフローチャートである。
【０１１９】
まず、指紋データ入力部２１によって採取記憶された多階調画像である照合指紋画像Ｂの全体が、回転変換処理部２４において回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）だけ回転変換される（ステップＳ１）。
【０１２０】
但し、この回転ずれ補正角度θｍが初期設定によりθ１＝０°のときだけ、つまりこのステップＳ１における照合指紋画像Ｂの最初の回転変換処理のときだけは、実際の回転変換は実施されない。
【０１２１】
すると、テンプレート配置処理部２５において、登録指紋データ記憶部２２にて記憶された非回転の登録指紋画像Ａに対し、複数の矩形テンプレート領域Ａ_ｉ［ｉ＝１〜Ｎ：Ｎ≧２の整数（この場合Ｎ＝５）］が定義されて配置される（ステップＳ２）。
【０１２２】
すると、照合処理部２３内の最大相関領域検出処理部２６において、前記登録指紋画像Ａに対し定義配置された各矩形テンプレート領域Ａｉに相当する部分画像が、前記回転変換処理部２４においてθｍ回転変換された（θ１＝０°のときは実質未回転の）照合指紋画像Ｂ上で水平方向及び垂直方向に１画素ずつラスター走査されながら、順次該テンプレートＡｉ内の全ての画素データとそれに対応する照合指紋画像Ｂ内の画素データとを用いた相関係数が計算される。そして、その相関係数が最大となる照合指紋画像Ｂ内の領域Ｂｉが検出される（ステップＳ３，Ｓ４）。
【０１２３】
ここで、登録指紋画像Ａ上に定義した第１の矩形テンプレート領域Ａ１に対し照合指紋画像Ｂから領域Ｂ１を検出する際に計算された最大相関係数値がＭＡＸＣＣとされテンプレート配置指紋データ選択処理部２９へ与えられる。
【０１２４】
この非回転の登録指紋画像Ａ上に定義した各矩形テンプレート領域Ａｉの画像データを基準とする回転処理後照合指紋画像Ｂ上での最大相関係数の算出及びその領域Ｂｉの検出は、各矩形テンプレート領域Ａｉ毎に順次行われ（ステップＳ３〜Ｓ５）、全ての矩形テンプレート領域Ａｉにそれぞれ最大の相関係数を有する照合指紋画像Ｂ上の各領域Ｂｉが検出されたと判断されると、照合処理部２３内の照合判定処理部２７において、登録指紋画像Ａ内に定義した前記複数の矩形テンプレート領域｛Ａ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）と照合指紋画像Ｂ上から検出した前記複数の領域｛Ｂ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）とが比較され、該登録指紋Ａと照合指紋Ｂの同一性が評価される（ステップＳ５→Ｓ６）。
【０１２５】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ７→終わり）。
【０１２６】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、前記回転変換処理部２４に現在設定されている回転ずれ補正角度θｍが初期設定に伴うθ１＝０°であるか判断される（ステップＳ７→Ｓ７ａ）。
【０１２７】
ここで、前記回転変換処理部２４に現在設定されている回転ずれ補正角度θｍが初期設定に伴うθ１＝０°であると判断された場合には、続いて、登録指紋データ記憶部２２に記憶されている登録指紋画像Ａの全体が、回転変換処理部２４において回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）だけ回転変換される（ステップＳ７ａ→Ｓ８ａ）。
【０１２８】
但し、この回転ずれ補正角度θｍが初期設定によりθ１＝０°のときだけ、つまりこのステップＳ８ａにおける登録指紋画像Ａの最初の回転変換処理のときだけは、実際の回転変換は実施されない。
【０１２９】
すると、テンプレート配置処理部２５において、前記指紋データ入力部２１により採取記憶された非回転の照合指紋画像Ｂに対し、複数の矩形テンプレート領域ｂ_ｉ［ｉ＝１〜Ｎ：Ｎ≧２の整数（この場合Ｎ＝５）］が定義されて配置される（ステップＳ８）。
【０１３０】
すると、照合処理部２３内の最大相関領域検出処理部２６において、前記非回転の照合指紋画像Ｂに対し定義配置された各矩形テンプレート領域ｂｉに相当する部分画像が、前記回転変換処理部２４によりθｍ回転変換された（θ１＝０°のときは実質未回転の）登録指紋画像Ａ上で水平方向及び垂直方向に１画素ずつラスター走査されながら、順次該テンプレートｂｉ内の全ての画素データとそれに対応する登録指紋画像Ａ内の画素データとを用いた相関係数が計算される。そして、その相関係数が最大となる登録指紋画像Ａ内の領域ａｉが検出される（ステップＳ９，Ｓ１０）。
【０１３１】
ここで、照合指紋画像Ｂ上に定義した第１の矩形テンプレート領域ｂ１に対し登録指紋画像Ａから領域ａ１を検出する際に計算された最大相関係数値がｍａｘｃｃとされテンプレート配置指紋データ選択処理部２９へ与えられる。
【０１３２】
この非回転の照合指紋画像Ｂ上に定義した各矩形テンプレート領域ｂｉの画像データを基準とする回転処理後登録指紋画像Ａ上での最大相関係数の算出及びその領域ａｉの検出は、各矩形テンプレート領域ｂｉ毎に順次行われ（ステップＳ９〜Ｓ１１）、全ての矩形テンプレート領域ｂｉにそれぞれ最大の相関係数を有する登録指紋画像Ａ上の各領域ａｉが検出されたと判断されると、照合処理部２３内の照合判定処理部２７において、照合指紋画像Ｂ内に定義した前記複数の矩形テンプレート領域｛ｂ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）と登録指紋画像Ａ上から検出した前記複数の領域｛ａ_ｉ｜ｉ＝１，２，…，Ｎ｝の分布（相互位置関係）とが比較され、該登録指紋Ａと照合指紋Ｂの同一性が評価される（ステップＳ１１→Ｓ１２）。
【０１３３】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ１３→終わり）。
【０１３４】
一方、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない場合には、前記回転変換処理部２４に現在設定されている回転ずれ補正角度θｍが初期設定に伴うθ１＝０°であるか判断される（ステップＳ１３→Ｓ１３ａ）。
【０１３５】
ここで、前記回転変換処理部２４に現在設定されている回転ずれ補正角度θｍが初期設定に伴うθ１＝０°であると判断された場合には、テンプレート配置指紋データ選択処理部２９において前記最大相関領域検出処理部２６から与えられた登録指紋画像Ａ上に配置した第１の矩形テンプレート領域Ａ１に対する照合指紋画像Ｂとの最大相関係数値ＭＡＸＣＣと、照合指紋画像Ｂ上に配置した第１の矩形テンプレート領域ｂ１に対する登録指紋画像Ａとの最大相関係数値ｍａｘｃｃとの大小が比較判断される（ステップＳ１３ａ→Ｓ１５）。
【０１３６】
そして、登録指紋画像Ａ上に配置した第１の矩形テンプレート領域Ａ１に対する照合指紋画像Ｂとの最大相関係数値ＭＡＸＣＣの方が、照合指紋画像Ｂ上に配置した第１の矩形テンプレート領域ｂ１に対する登録指紋画像Ａとの最大相関係数値ｍａｘｃｃよりも大きい（あるいは以上）と判断された場合には、前記回転変換処理部２４における回転ずれ補正角度θｍが順次更新設定されるθ２＝１°以降の処理において、登録指紋画像Ａにテンプレート領域Ａｉを定義配置し照合指紋画像Ｂを順次回転変換して指紋照合を行う前記ステップＳ１〜Ｓ７での指紋照合処理が実施される（ステップＳ１５→Ｓ１〜Ｓ７）。
【０１３７】
そして、このステップＳ７において、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない状態では、前記回転変換処理部２４において規定した全ての回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）により回転変換された各回転後照合指紋画像Ｂのそれぞれについて、非回転の登録指紋画像Ａにテンプレート領域Ａｉを配置した照合処理が実施されたか判断される（ステップＳ７→Ｓ１４ａ）。
【０１３８】
そして、前記規定した全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについて、非回転の登録指紋画像Ａにテンプレート領域Ａｉを配置した照合処理が実施されてないと判断された場合には、再びステップＳ１からの処理に戻り、更新した回転ずれ補正角度θｍ（＝ｍ＋１）により回転変換された照合指紋画像Ｂについての非回転の登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置した照合処理が繰り返し実施される（ステップＳ１４ａ→Ｓ１〜Ｓ７）。
【０１３９】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ７→終わり）。
【０１４０】
また、全ての回転ずれ補正角度θｍにより回転変換された各回転後の照合指紋画像Ｂについての非回転の登録指紋画像Ａとの照合処理が実施されたと判断された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋でないことが表示される（ステップＳ１４ａ→終わり）。
【０１４１】
一方、前記ステップＳ１５において、照合指紋画像Ｂ上に配置した第１の矩形テンプレート領域ｂ１に対する登録指紋画像Ａとの最大相関係数値ｍａｘｃｃの方が、登録指紋画像Ａ上に配置した第１の矩形テンプレート領域Ａ１に対する照合指紋画像Ｂとの最大相関係数値ＭＡＸＣＣよりも大きいと判断された場合には、前記回転変換処理部２４における回転ずれ補正角度θｍが順次更新設定されるθ２＝１°以降の処理において、照合指紋画像Ｂにテンプレート領域ｂｉを定義配置し登録指紋画像Ａを順次回転変換して指紋照合を行う前記ステップＳ８ａ〜Ｓ１３での指紋照合処理が実施される（ステップＳ１５→Ｓ８ａ〜Ｓ１３）。
【０１４２】
そして、このステップＳ１３において、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定されない状態では、前記回転変換処理部２４において規定した全ての回転ずれ補正角度θｍ（＝０°，１°，−１°，２°，−２°，３°，−３°，…，ｋ°，−ｋ°［ｍ＝１〜ｋ×２＋１、ｋ≧１の整数］）により回転変換された各回転後登録指紋画像Ａのそれぞれについて、非回転の照合指紋画像Ｂにテンプレート領域ｂｉを配置した照合処理が実施されたか判断される（ステップＳ１３→Ｓ１４ｂ）。
【０１４３】
そして、前記規定した全ての回転ずれ補正角度θｍにより回転変換された各回転後の登録指紋画像Ａについて、非回転の照合指紋画像Ｂにテンプレート領域ｂｉを配置した照合処理が実施されてないと判断された場合には、再びステップＳ８ａからの処理に戻り、更新した回転ずれ補正角度θｍ（＝ｍ＋１）により回転変換された登録指紋画像Ａについての非回転の照合指紋画像Ｂ上にテンプレート領域ｂｉを定義配置した照合処理が繰り返し実施される（ステップＳ１４ｂ→Ｓ８ａ〜Ｓ１３）。
【０１４４】
そして、登録指紋Ａと照合指紋Ｂとが同一指紋であると判定された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋であることが表示される（ステップＳ１３→終わり）。
【０１４５】
また、全ての回転ずれ補正角度θｍにより回転変換された各回転後の登録指紋画像Ａについての非回転の照合指紋画像Ｂとの照合処理が実施されたと判断された場合には、前記一連の照合処理は終了され、照合結果表示部２８（１５）により登録指紋Ａと照合指紋Ｂとが同一指紋でないことが表示される（ステップＳ１４ｂ→終わり）。
【０１４６】
したがって、この第４実施形態の指紋データ照合装置によれば、登録指紋画像Ａにテンプレート領域Ａｉを定義配置して順次回転変換される照合指紋画像Ｂとの照合を行う指紋照合処理と、照合指紋画像Ｂにテンプレート領域ｂｉを定義配置して順次回転変換される登録指紋画像Ａとの照合を行う指紋照合処理とで、各初期ルーチンで大きな相関係数値が得られる方の指紋照合処理に固定して処理を実施する構成としたので、前記第１乃至第３実施形態と同様に、回転ずれを有する同一人物・同一指の指紋照合を行うのに、より小さな回転ずれ補正角度θｍにおいて照合一致が得られる可能性が高くなるだけでなく、さらに少ない演算処理量にして照合時間を大幅短縮できるようになる。
【０１４７】
なお、この第４実施形態の指紋照合処理におけるステップＳ１５では、登録指紋画像Ａ上に配置した第１の矩形テンプレート領域Ａ１に対する照合指紋画像Ｂとの最大相関係数値ＭＡＸＣＣと、照合指紋画像Ｂ上に配置した第１の矩形テンプレート領域ｂ１に対する登録指紋画像Ａとの最大相関係数値ｍａｘｃｃとの大小を比較判断する構成としたが、これに代わり、登録指紋画像Ａ上に配置した複数の矩形テンプレート領域Ａｉのそれぞれに対する照合指紋画像Ｂとの各最大相関係数値ＭＡＸＣＣの平均値と、照合指紋画像Ｂ上に配置した複数の矩形テンプレート領域ｂｉのそれぞれに対する登録指紋画像Ａとの各最大相関係数値ｍａｘｃｃの平均値との大小を比較判断する構成としてもよい。
【０１４８】
また、前記第４実施形態では、照合指紋画像Ｂを回転変換したときには非回転の登録指紋画像Ａ上にテンプレート領域Ａｉを定義配置して該回転処理後の照合指紋画像Ｂとの照合処理を行い、登録指紋画像Ａを回転変換したときには非回転の照合指紋画像Ｂ上にテンプレートｂｉを定義配置して該回転処理後の登録指紋画像Ａとの照合処理を行う構成としたが、次の第５実施形態において説明するように、回転変換した照合指紋画像Ｂあるいは登録指紋画像Ａに対してテンプレート領域ｂｉ、Ａｉを定義配置し、非回転の登録指紋画像Ａとの照合処理あるいは非回転の照合指紋画像Ｂとの照合処理を行う構成としてもよい。
【０１４９】
（第５実施形態）
図１７は前記指紋データ照合装置の第５実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図である。
【０１５０】
この第５実施形態の指紋照合動作機能では、テンプレート配置指紋データ選択処理部２９を通して登録指紋画像Ａを回転変換処理部２４で回転変換し、この回転変換した登録指紋画像Ａに対しテンプレート配置処理部２５でテンプレート領域Ａｉを定義配置した場合には、前記テンプレート配置指紋データ選択処理部２９を通して非回転の照合指紋画像Ｂを直接最大相関領域検出処理部２６へ与え、また、テンプレート配置指紋データ選択処理部２９を通して照合指紋画像Ｂを回転変換処理部２４で回転変換し、この回転変換した照合指紋画像Ｂに対しテンプレート配置処理部２５でテンプレート領域ｂｉを定義配置した場合には、前記テンプレート配置指紋データ選択処理部２９を通して非回転の登録指紋画像Ａを直接最大相関領域検出処理部２６へ与える。
【０１５１】
そして、初期設定された回転ずれ補正角度θｍ（θ１＝０°）において、登録指紋画像Ａを回転変換しテンプレート領域Ａｉを定義配置して非回転の照合指紋画像Ｂとの間で計算される最大相関係数値ＭＡＸＣＣと、同初期設定された回転ずれ補正角度θｍ（θ１＝０°）において、照合指紋画像Ｂを回転変換しテンプレート領域ｂｉを定義配置して非回転の登録指紋画像Ａとの間で計算される最大相関係数値ｍａｘｃｃとを、最大相関領域検出処理部２６からテンプレート配置指紋データ選択処理部２９へ与えて大小比較判断し、順次更新される回転ずれ補正角度θｍであるθ２＝１°以降の処理において、前記登録指紋画像Ａを回転変換しテンプレート領域Ａｉを定義配置して非回転の照合指紋画像Ｂとの間で行う指紋照合処理か、あるいは前記照合指紋画像Ｂを回転変換しテンプレート領域ｂｉを定義配置して非回転の登録指紋画像Ａとの間で行う指紋照合処理かを選択決定する構成とする。
【０１５２】
したがって、この第５実施形態の指紋データ照合装置によっても、前記第４実施形態と同様に、回転ずれを有する同一人物・同一指の指紋照合を行うのに、より小さな回転ずれ補正角度θｍにおいて照合一致が得られる可能性が高くなるだけでなく、さらに少ない演算処理量にして照合時間を大幅短縮できるようになる。
【０１５３】
なお、前記第３実施形態及び第５実施形態において、テンプレート領域を定義配置する側の登録指紋画像Ａあるいは照合指紋画像Ｂの回転変換処理では、該当する指紋画像ＡあるいはＢの全体を回転変換する構成としたが、指紋画像ＡあるいはＢのテンプレート領域の定義配置の対象となる部分的な指紋画像領域のみを回転変換する構成としてもよい。これによれば、回転変換処理を掛ける画像量そのものを削減でき、この回転変換処理時間を短縮して指紋照合処理時間をさらに短縮できるようになる。
【０１５４】
また、前記各実施形態における登録指紋画像Ａと照合指紋画像Ｂとの照合処理では、Ａｉ、Ｂｉ、ｂｉ、ａｉの各領域を矩形領域としていたが、これらの領域の形状は矩形に限らず任意でよい。また、それら領域Ａｉ、Ｂｉ、ｂｉ、ａｉの大きさ・形状は同一であることが望ましいが、多少の違いがあったとしても、要求される指紋照合の照合精度が得られる限り許容できる。さらに、各テンプレート領域Ａｉ、ｂｉを定義して配置する位置も任意でよい。
【０１５５】
また、照合判定においては、前述した処理におけるΔｉを評価する方法の他に、例えばＡｉ又はｂｉを頂点として形成される図形とＢｉ又はａｉを頂点として形成される図形との、形状、あるいは面積の違いに基づく判定など、様々な方法を採用することも可能である。
【０１５６】
なお、前記各実施形態において記載した手法、すなわち、図３〜図５及び図１０〜図１７における各動作機能ブロック図や各フローチャートに基づく第１乃至第５実施形態のそれぞれにおける指紋照合処理などの各手法は、コンピュータに実行させることができるプログラムとして、メモリカード（ＲＯＭカード、ＲＡＭカード等）、磁気ディスク（フロッピディスク、ハードディスク等）、光ディスク（ＣＤ−ＲＯＭ、ＤＶＤ等）、半導体メモリ等の外部記憶媒体１８ａに格納して配布することができる。そして、指紋データ照合装置のコンピュータは、この外部記憶媒体１８ａに記憶されたプログラムを記憶装置１２に読み込み、この読み込んだプログラムによって動作が制御されることにより、前記実施形態において説明した指紋照合機能を実現し、前述した手法による同様の処理を実行することができる。
【０１５７】
また、前記各手法を実現するためのプログラムのデータは、プログラムコードの形態としてネットワーク（公衆回線）２０上を伝送させることができ、このネットワーク２０に接続された指紋データ照合装置の伝送制御部１９によって前記のプログラムデータを取り込み、前述した指紋照合機能を実現することもできる。
【０１５８】
【発明の効果】
以上のように、本発明によれば、画像データＡと画像データＢとを照合する際に、画像を回転させると共に、テンプレートを画像Ａと画像Ｂのそれぞれに交互に配置する２通りの照合処理を行うので、一方にテンプレートを配置した照合処理で一致が得られなくても、他方にテンプレートを配置した照合処理で一致を得る可能性があり、画像データの回転処理回数が少ないうちに画像ＡとＢの照合一致が図れる可能性が高くなり照合時間を短縮できるようになる。
【図面の簡単な説明】
【図１】本発明の画像データ照合装置の実施形態に係る指紋データ照合装置の電子回路の構成を示すブロック図。
【図２】前記指紋データ照合装置のＲＡＭに確保されるデータメモリを示す図。
【図３】前記指紋データ照合装置の第１実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図。
【図４】前記指紋データ照合装置の第１実施形態の指紋照合処理（その１）を示すフローチャート。
【図５】前記指紋データ照合装置の第１実施形態の指紋照合処理（その２）を示すフローチャート。
【図６】前記指紋データ照合装置の指紋照合処理に伴う登録指紋画像Ａ上でのテンプレート領域Ａｉの配置位置分布と照合指紋画像Ｂ上での最大相関領域Ｂｉの検出位置分布との対比状態を示す図。
【図７】前記指紋データ照合装置の指紋照合処理に伴いＲＡＭ内の登録指紋位置メモリと照合指紋位置メモリとにそれぞれ格納されるテンプレート位置と最大相関領域検出位置との各位置データを示す図。
【図８】前記指紋データ照合装置の指紋照合処理に伴う照合指紋画像Ｂ上でのテンプレート領域ｂｉの配置位置分布と登録指紋画像Ａ上での最大相関領域ａｉの検出位置分布との対比状態を示す図。
【図９】前記指紋データ照合装置の指紋照合処理に伴いＲＡＭ内の照合指紋位置メモリと登録指紋位置メモリとにそれぞれ格納されるテンプレート位置と最大相関領域検出位置との各位置データを示す図。
【図１０】前記指紋データ照合装置の第２実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図。
【図１１】前記指紋データ照合装置の第２実施形態の指紋照合処理（その１）を示すフローチャート。
【図１２】前記指紋データ照合装置の第２実施形態の指紋照合処理（その２）を示すフローチャート。
【図１３】前記指紋データ照合装置の第３実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図。
【図１４】前記指紋データ照合装置の第４実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図。
【図１５】前記指紋データ照合装置の第４実施形態の指紋照合処理（その１）を示すフローチャート。
【図１６】前記指紋データ照合装置の第４実施形態の指紋照合処理（その２）を示すフローチャート。
【図１７】前記指紋データ照合装置の第５実施形態の指紋照合処理プログラムの実行に伴う動作機能の構成を示すブロック図。
【図１８】先願の指紋データ照合装置における指紋データ照合の手順を示すフローチャート。
【図１９】前記先願の指紋データ照合手順に伴い登録指紋データ内に配置された複数の基準領域及びこれに基づき照合指紋データ上で検出された複数の領域を示す図。
【図２０】前記先願の指紋データ照合手順に伴い登録指紋データ内に配置された複数の基準領域及びこれに基づき照合指紋データ上で不適当な複数の領域が検出された場合の例を示す図。
【符号の説明】
１１ …制御部（ＣＰＵ）
１２ …記憶装置
１３ …ＲＡＭ
１３ａ…読み取り画像メモリ
１３ｂ…照合指紋画像メモリ
１３ｃ…照合指紋回転画像メモリ
１３ｄ…登録指紋位置メモリ
１３ｅ…照合指紋位置メモリ
１３ｆ…回転角メモリ
１４ …イメージ読み取り装置
１５ …表示部
１６ …入力部
１７ …バス
１８ …記憶媒体読み取り部
１８ａ…外部記憶媒体
１９ …伝送制御部
２０ …ネットワーク
２０ａ…プログラムサーバ（外部のコンピュータ端末）
２０ｂ…外部端末の記憶装置
２１ …指紋データ入力部
２２ …登録指紋データ記憶部
２３ …照合処理部
２４ …回転変換処理部
２５ …テンプレート配置処理部
２６ …最大相関領域検出処理部
２７ …照合判定処理部
２８ …照合結果表示部
２９ …テンプレート配置指紋データ選択処理部[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to, for example, an image data collation device for collating fingerprint image data, an image data collation method, and a storage medium storing an image data collation processing program.
[0002]
[Prior art]
For example, as one of the conventional fingerprint collation functions operating on a personal computer (PC), the fingerprint data extracted from two fingerprint data of registered fingerprint data (registered fingerprint data) and fingerprint data to be collated (collation fingerprint data). There is a fingerprint matching method “feature figure matching method” for determining the identity of the two fingerprint data based on the positional relationship between characteristic figures (branch points, end points, etc.). The “characteristic figure matching method” is effective in many cases where images are compared with each other in addition to fingerprints, and its effectiveness is well known.
[0003]
However, the fingerprint data of the subject to be registered or collated
(1) The “valley” of the fingerprint is shallow, and the fingerprint data is not clear.
[0004]
(2) The skin is soft, and the “mountain” of the fingerprint is easily broken.
[0005]
(3) The position where the characteristic figure (branch point, end point, etc.) appears is likely to change.
[0006]
(4) If the number of characteristic figures (branch points, end points, etc.) has a characteristic that changes greatly every time fingerprint data is collected, the above-mentioned "characteristic figure collation method" increases the probability of failing the collation. It is said that there are about two to three persons per 100 persons who have drawbacks and have fingerprints having such characteristics.
[0007]
This is too large in consideration of the fact that a standard value of about 0.1% is set as the false rejection rate or false acceptance rate at the time of fingerprint data collation, which is a problem in authentication accuracy.
[0008]
Therefore, as a fingerprint data matching method that solves the problem of the "characteristic figure matching method", an image data matching method filed by the same applicant (Japanese Patent Application No. 10-372205: filed on December 28, 1998) ).
[0009]
In the image data collation method of the prior application, even in the case of fingerprint data having the above-described characteristics, it is possible to suppress the probability of collation failure and to collate the registered fingerprint data with the collation fingerprint data with sufficient authentication accuracy. A fingerprint data matching function is disclosed.
[0010]
FIG. 18 is a flowchart showing the procedure of fingerprint data collation in the fingerprint data collation apparatus of the prior application.
[0011]
FIG. 19 is a diagram showing a plurality of reference areas arranged in the registered fingerprint data in accordance with the fingerprint data collation procedure of the prior application and a plurality of areas detected on the collation fingerprint data based on the reference areas.
[0012]
That is, as shown in FIGS. 18 and 19,
(1) Multi-tone registered fingerprint data: a plurality of rectangular areas A in A_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged (see FIG. 19A).
[0013]
(2) Each rectangular area A_iIs scanned on the collation fingerprint data: B, and the respective rectangular areas A are scanned._iThe correlation coefficient is calculated using the pixel data in the reference fingerprint data B and the corresponding pixel data in the reference fingerprint data B.
[0014]
(3) a plurality of regions B in which the correlation coefficient is maximum_iIs detected from the collation fingerprint data B (see FIG. 19B).
[0015]
(4) The plurality of rectangular areas ｛A defined in the registered fingerprint data A_i| I = 1, 2,..., N} and the plurality of areas {B detected from the verification fingerprint data B_i| I = 1, 2,..., N} and evaluates the identity of the registered fingerprint A and the collated fingerprint B. This is a fingerprint data collation method “template collation method using correlation coefficient”.
[0016]
As described above, the “template matching method based on correlation coefficient” which is the fingerprint data matching method of the prior application is based on each area A in which a plurality of areas are defined and registered in the registered fingerprint data A._iIn contrast, an area B where the correlation coefficient is maximum using pixel data that can exist in the collation fingerprint data B_iIs detected from the collation fingerprint data B, and the area ΔA defined in the registered fingerprint data A is detected._i| I = 1, 2,..., N} and the area {B detected from collation fingerprint data B_i| I = 1, 2,..., N} and evaluates the identity of the registered fingerprint data A and the collated fingerprint data B for collation. 20 is not taken into account. For example, as shown in FIG. 20, when collecting the collation fingerprint data B, the collation fingerprint data B In the case where the collation fingerprint data is the same person and the same fingerprint data as the registered fingerprint data, the area B where the correlation coefficient is the maximum is obtained even if the collation fingerprint data is the same fingerprint data._iIs detected as an inappropriate area, and the above defined area ｛A_i| I = 1, 2,..., N} and the detection area {B_i| I = 1, 2,..., N} are different, and are evaluated as not matching the registered fingerprint data.
[0017]
FIG. 20 is a diagram showing an example in which a plurality of reference areas arranged in registered fingerprint data in accordance with the fingerprint data collation procedure of the prior application and a plurality of inappropriate areas based on the reference areas are detected on the collation fingerprint data. It is.
[0018]
[Problems to be solved by the invention]
SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems, and enables an image that can be accurately compared with registered image data and collation image data even in a state where a rotation error occurs in a collation image. It is an object of the present invention to provide a data collation device, an image data collation method, and a storage medium storing an image data collation processing program.
[0019]
[Means for Solving the Problems]
Of the present inventionFirst perspectiveIs an image data collating apparatus that collates image data A and image data B to determine the identity of the two, and the image data A or the image data BOne ofRotating means for rotating the image data A or B by a predetermined angle, and first template arranging means for arranging a template defining a plurality of regions with respect to the image data A every time the image data A or B is rotated by the image rotating means. A first correlation area detecting means for detecting, on the image data B, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data A by the first template arranging means; Comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the first template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting unit A first collation judging unit for judging identity between the image data A and the image data B, and each time the image data A or B is rotated by the image rotating unit. Second template arranging means for arranging a template defining a plurality of areas with respect to the image data B, and a maximum correlation with each of the plurality of areas arranged in the image data B by the second template arranging means A second correlation area detecting means for detecting a plurality of correlation areas on the image data A; a mutual positional relationship between the plurality of areas arranged in the image data B by the second template arranging means; and the second correlation area A second comparison / judgment unit for comparing the mutual positional relationship between the plurality of correlation regions detected on the image data A by the detection unit and judging the identity between the image data A and the image data B. It is characterized by the following.
[0020]
AlsoOf the present inventionSecond perspectiveImage data collating device according toIs an image data collating apparatus for collating image data A and image data B to determine the identity between the two, and an image rotating means for rotating the image data A and the image data B alternately by a preset angle And a template for defining a plurality of regions with respect to the non-rotated image data B when the image data A is rotated by the image rotating means, and the non-rotated image data is set when the image data B is rotated. A template arranging means for arranging a template defining a plurality of areas with respect to A; and when the template arranging means arranges a template in the image data B, each of the plurality of areas defined by the template is First correlation area detecting means for detecting a plurality of correlation areas having a correlation on image data A, and the template arranging means By comparing the mutual positional relationship between the plurality of regions arranged in the image data B with the mutual positional relationship between the plurality of correlation regions detected on the image data A by the first correlation region detecting means, the image data A is compared. A first collation judging unit for judging the identity of the image data B with the image data B, and when a template is arranged on the image data A by the template arranging unit, the maximum correlation with each of a plurality of regions defined by the template A second correlation area detecting means for detecting a plurality of correlation areas having a relationship on the image data B, a mutual positional relationship between the plurality of areas arranged in the image data A by the template arranging means, and the second correlation area The identity of the image data A and the image data B is determined by comparing the mutual positional relationship between the plurality of correlation areas detected on the image data B by the detection means. A second match determination means that, characterized by comprising a.
[0021]
Of the present inventionThird perspectiveThe image data matching device according toAn image data collation device that collates image data A and image data B to determine the identity of the two, and an image rotation unit that alternately rotates the image data A and the image data B by a preset angle, When the image data A is rotated by the image rotating means, a template defining a plurality of regions is arranged for the image data A, and when the image data B is rotated, a plurality of templates are defined for the image data B. A template arranging means for arranging a template defining an area, and a plurality of areas having the maximum correlation with each of the plurality of areas defined by the template when the template arranging means arranges the template in the image data A. A first correlation area detecting means for detecting the correlation area on the non-rotated image data B; And comparing the mutual positional relationship of the plurality of regions arranged in the image data A with the mutual positional relationship of the plurality of correlation regions detected on the image data B by the first correlation region detecting means. First matching determining means for determining the identity of the image data B with a plurality of areas defined by the template when a template is arranged on the image data B by the template arranging means. A second correlation area detecting means for detecting a plurality of correlation areas having a relationship on the non-rotated image data A; a mutual positional relationship between a plurality of areas arranged in the image data B by the template arranging means; And comparing the image data A with the image data B by comparing the mutual positional relationship of the plurality of correlation regions detected on the image data A by the correlation region detecting means. A second matching determination means for determining,It is characterized by having.
[0023]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0024]
(1st Embodiment)
FIG. 1 is a block diagram showing a configuration of an electronic circuit of a fingerprint data matching device according to an embodiment of the image data matching device of the present invention.
[0025]
This fingerprint data collation device includes a control unit (CPU) 11, a storage device 12, a RAM 13, an image reading device 14, a display unit 15, an input unit 16, a storage medium reading unit 18, and a transmission control unit 19. Connected to each other.
[0026]
The control unit (CPU) 11 responds to an input signal from the input unit 12 with a control program stored in advance in the storage device 12 or a storage medium such as a floppy disk drive FDD from an external storage medium 18a such as a floppy disk FD. The fingerprint collation processing program read into the storage device 12 via the reading unit 18 or from the storage device 20b of the external computer terminal (program server 20a) via the communication network 20 to the storage device 12 via the transmission control unit 19 The read fingerprint matching program is started, and the operation of each part of the circuit is controlled using the RAM 13 as a work memory.
[0027]
The storage device 12 is configured by a ROM of a semiconductor memory, a hard disk device HD, or the like. In addition to storing the above-described control program read by the control unit (CPU) 11 immediately after turning on the power of the present verification device, the storage device 12 also has an image reading function. The registered fingerprint image of the subject read in advance by the device 14 is stored.
[0028]
The RAM 13 is a work memory used by the control unit (CPU) 11 when executing the above-described control program.
[0029]
The image reading device 14 is a device for acquiring (collecting) a registered fingerprint image or a collation fingerprint image of a subject, and is, for example, an image scanner device or an image sensor such as a CCD (charge coupled device).
[0030]
The display unit 15 is a display device, such as a CRT or a liquid crystal, for displaying a determination result of collation between the registered fingerprint image and the collation fingerprint image.
[0031]
The input unit 16 is an input device, such as a keyboard device, for the user of the present collation device to instruct the control unit (CPU) 11 to acquire a fingerprint image or to start a fingerprint collation process described later.
[0032]
FIG. 2 is a diagram showing a data memory secured in the RAM 13 of the fingerprint data collating apparatus.
[0033]
In the RAM 13, a read image memory 13a, a collation fingerprint image memory 13b, a collation fingerprint rotation image memory 13c, a registered fingerprint position memory 13d, a collation fingerprint position memory 13e, a rotation angle memory 13f, and the like are secured.
[0034]
In the read image memory 13a, the image data of the fingerprint of the subject read by the image reading device 14 is temporarily stored as a multi-tone image.
[0035]
The collation fingerprint image memory 13b stores the image data of the fingerprint of the subject to be collated read by the image reader 14 as a multi-tone image, and the collation fingerprint image data is stored in the storage device 12. The registered fingerprint image data is normalized and stored in accordance with the number of pixels and the pixel pitch of the registered fingerprint image data.
[0036]
The collation fingerprint rotation image memory 13c stores the image data after the rotation process when the collation fingerprint image data stored in the collation fingerprint image memory 13b is rotated by the rotation deviation correction angle θm.
[0037]
In the registered fingerprint position memory 13d, coordinate data indicating the defined position of each of a plurality of rectangular areas Ai defined as templates with respect to the registered fingerprint image data A stored and registered in the storage device 12, or the reference fingerprint image memory 13b, when a plurality of rectangular areas bi are defined as templates for the collation fingerprint image data B stored in the registered fingerprint image data B, the respective partial images and the maximum correlation coefficient are calculated. Are detected, and coordinate data indicating the positions of the plurality of detected rectangular areas ai on the registered fingerprint image data A are stored.
[0038]
When a plurality of rectangular areas Ai are defined as templates for the registered fingerprint image data A stored and registered in the storage device 12, a maximum correlation coefficient with each partial image is calculated in the verification fingerprint position memory 13e. A plurality of rectangular areas Bi on the collated fingerprint image data B to be detected are detected, and coordinate data indicating the positions of the plurality of detected rectangular areas Bi on the collated fingerprint image data B, or the collated fingerprint image memory Coordinate data indicating the defined position of each of the plurality of rectangular regions bi defined as templates for the collation fingerprint image data B stored in 13b is stored.
[0039]
The rotation angle memory 13h stores the current rotation angle θm of the verification fingerprint image data B stored in the verification fingerprint rotation image memory 13c.
[0040]
FIG. 3 is a block diagram showing a configuration of an operation function associated with execution of the fingerprint collation processing program of the first embodiment of the fingerprint data collation apparatus.
[0041]
In the functional blocks of the operation function in the fingerprint data collation apparatus, the corresponding components of the electronic circuit in FIGS. 1 and 2 are indicated by reference numerals in parentheses.
[0042]
The fingerprint data input unit 21 includes an image reading device 14, a read image memory 13a, and a collation fingerprint image memory 13b, and inputs a fingerprint image to be registered or a fingerprint image to be collated as a multi-tone image. The registered fingerprint image data A input by the fingerprint data input unit 21 is provided to the registered fingerprint data storage unit 22 and stored and registered in the storage device 12, and the collation fingerprint image data B is subjected to rotation conversion processing in the collation processing unit 23. Section 24.
[0043]
The registered fingerprint image data A stored and registered by the registered fingerprint data storage unit 22 is provided to the template arrangement processing unit 25 or the maximum correlation area detection processing unit 26 in the matching processing unit 23.
[0044]
The verification processing unit 23 includes a control unit (CPU) 11, a fingerprint verification processing program in the storage device 12, a verification fingerprint rotation image memory 13c, a registered fingerprint position memory 13d, a verification fingerprint position memory 13e, and a rotation angle memory 13f. .
[0045]
The rotation conversion processing unit 24 converts the entire collation fingerprint image data B provided from the fingerprint data input unit 21 into an angle θm (= 0, 1, −1, 2, −2, 3, −3,..., K, − k (m = 1 to k × 2 + 1, k ≧ 1 integer)) (the actual rotation conversion is not performed when θ1 = 0). The rotation-converted collation fingerprint image data B is provided to the maximum correlation area detection processing unit 26 or the template arrangement processing unit 25.
[0046]
The template arrangement processing unit 25 stores a plurality of rectangular areas A in the registered fingerprint image data A stored and registered in the registered fingerprint data storage unit 22._i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged as a template (see FIG. 6A), or rotated by an angle θm by the rotation conversion processing unit 24. For the converted fingerprint image data B, a plurality of rectangular areas b_i[I = 1 to N: an integer of N ≧ 2 (in this case, N = 5)] is defined and arranged as a template (see FIG. 8B). The partial image data in the range corresponding to each of the plurality of rectangular regions Ai arranged on A or the partial image data in the range corresponding to each of the plurality of rectangular regions bi arranged on the verification fingerprint image data B are: It is provided to the maximum correlation area detection processing unit 26.
[0047]
When a plurality of rectangular areas Ai are arranged with respect to the registered fingerprint image data A in the template arrangement processing section 25, the maximum correlation area detection processing section 26 performs rotation conversion on the partial images corresponding to each of the rectangular areas Ai. While scanning on the collation fingerprint image data B rotated and converted by the angle θm in the processing unit 24, the rectangular area A_iThe correlation coefficient is calculated using the pixel data in the area and the corresponding pixel data in the collation fingerprint image data B, and a plurality of areas (maximum correlation areas) B having the maximum correlation coefficient are calculated._iIs detected from the collation fingerprint image data B (see FIG. 6), or when a plurality of rectangular areas bi are arranged in the collation fingerprint image data B after the rotation conversion in the template While scanning the partial image corresponding to each area bi on the registered fingerprint image data A, the respective rectangular areas b_iA correlation coefficient is calculated using the pixel data in the pixel data and the corresponding pixel data in the registered fingerprint image data A, and a plurality of areas (maximum correlation area) a where the correlation coefficient is maximum_iIs detected from the registered fingerprint image data A (see FIG. 8).
[0048]
When a plurality of rectangular areas Ai are arranged for the registered fingerprint image data A in the template arrangement processing unit 25, coordinate data indicating the position of each of the rectangular areas Ai on the registered fingerprint image data A, The coordinate data indicating the position of each of the plurality of maximum correlation areas Bi detected on the collation fingerprint image data B after the rotation conversion by the correlation area detection processing section 26 is given to the collation determination processing section 27, and the template When a plurality of rectangular areas bi are arranged in the collation fingerprint image data B after the rotation conversion in the arrangement processing unit 25, the coordinate data indicating the position of each of the rectangular areas bi on the collation fingerprint image data B and the maximum Coordinate data indicating the position of each of the plurality of maximum correlation areas ai detected on the registered fingerprint image data A by the correlation area detection processing unit 26 But given the collation determining unit 27.
[0049]
When a plurality of rectangular areas Ai are arranged with respect to the registered fingerprint image data A in the template arrangement processing section 25, the collation determination processing section 27 performs a process on the registered fingerprint image data A given from the template arrangement processing section 25. In the registered fingerprint image data A on the basis of the coordinate data indicating the position of each of the rectangular areas Ai and the coordinate data indicating the position of each of the plurality of maximum correlation areas Bi detected by the maximum correlation area detection processing unit 26. The arrangement position distribution (mutual positional relationship) of the area Ai is compared with the position distribution (mutual positional relationship) of the detection area Bi in the rotated fingerprint image data B. Based on the difference, the similarity between the registered fingerprint image and the verification fingerprint image is determined. To determine the identity. When a plurality of rectangular areas bi are arranged on the post-rotation verification fingerprint image data B in the template arrangement processing unit 25, on the post-rotation verification fingerprint image data B given from the template arrangement processing unit 25, Based on the coordinate data indicating the position of each of the rectangular areas bi and the coordinate data indicating the position of each of the plurality of maximum correlation areas ai detected by the maximum correlation area detection processing unit 26. The arrangement position distribution (mutual positional relationship) of the rectangular area bi is compared with the position distribution (mutual positional relationship) of the detection area ai in the registered fingerprint image data A, and the similarity between the registered fingerprint image and the verification fingerprint image is determined from the difference. Judge and evaluate identity. Then, the evaluation result of the identity between the registered fingerprint image and the collation fingerprint image obtained by the collation determination processing unit 27 is given to the collation result display unit 28 and displayed.
[0050]
Next, the fingerprint matching function of the first embodiment in the fingerprint data matching device having the above configuration will be described.
[0051]
FIG. 4 is a flowchart showing a fingerprint matching process (1) of the first embodiment of the fingerprint data matching device.
[0052]
FIG. 5 is a flowchart showing a fingerprint matching process (part 2) of the first embodiment of the fingerprint data matching device.
[0053]
FIG. 6 shows a comparison state between the arrangement position distribution of the template area Ai on the registered fingerprint image A and the detection position distribution of the maximum correlation area Bi on the collation fingerprint image B accompanying the fingerprint collation processing of the fingerprint data collation apparatus. FIG.
[0054]
FIG. 7 is a diagram showing respective position data of the template position and the maximum correlation area detection position stored in the registered fingerprint position memory 13d and the collated fingerprint position memory 13e in the RAM 13 with the fingerprint collation processing of the fingerprint data collation device. It is.
[0055]
FIG. 8 shows a comparison state between the arrangement position distribution of the template area bi on the collation fingerprint image B and the detection position distribution of the maximum correlation area ai on the registered fingerprint image A in the fingerprint collation processing of the fingerprint data collation apparatus. FIG.
[0056]
FIG. 9 is a diagram showing respective position data of the template position and the maximum correlation area detection position stored in the collation fingerprint position memory 13e and the registered fingerprint position memory 13d in the RAM 13 with the fingerprint collation processing of the fingerprint data collation device. It is.
[0057]
The fingerprint image A of the subject to be registered read by the image reading device 14 in the fingerprint data input unit 21 is stored in the storage device 12 in the registered fingerprint data storage unit 22 in advance and registered. When the fingerprint image of the subject is read by the image reading device 14, the fingerprint image is written to the read image memory 13a in the RAM 13 and transferred to the collation fingerprint image memory 13b for storage.
[0058]
Then, the fingerprint collation processing in FIGS. 4 and 5 is started, and first, the entire collation fingerprint image B which is a multi-tone image collected and stored in the collation fingerprint image memory 13b by the fingerprint data input unit 21 is subjected to the rotation conversion processing. The rotation deviation correction angle θm (= 0 °, 1 °, −1 °, 2 °, −2 °, 3 °, −3 °,..., K °, −k ° [m = 1 to k × 2 + 1] , K ≧ 1]] and is stored in the collation fingerprint rotation image memory 13c (step S1). However, when θ1 = 0, the actual rotation conversion is not performed.
[0059]
Then, in the template arrangement processing unit 25, as shown in FIG. 6A, a plurality of rectangular template areas A are registered with respect to the multi-tone registered fingerprint image A stored in the storage device 12 in the registered fingerprint data storage unit 22._i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged, and coordinate data indicating the position of each template region Ai is registered as shown in FIG. The data is stored in the fingerprint position memory 13d (step S2).
[0060]
Then, in the maximum correlation area detection processing unit 26 in the matching processing unit 23, the partial image corresponding to each rectangular template area Ai defined and arranged for the registered fingerprint image A is rotated by θm in the rotation conversion processing unit 24. While raster scanning is performed one pixel at a time in the horizontal and vertical directions on the collation fingerprint image B stored in the collation fingerprint rotation image memory 13c, all pixel data in the template Ai and the corresponding A correlation coefficient using the pixel data is calculated. Then, an area Bi in the collation fingerprint image B having the maximum correlation coefficient is detected as shown in FIG. 6B, and as shown in FIG. 7B, the detection of each maximum correlation area Bi is performed. The coordinate data indicating the position is stored in the collation fingerprint position memory 13e (steps S3 and S4). The calculation of the correlation coefficient will be described later.
[0061]
The calculation of the maximum correlation coefficient on the post-rotation verification fingerprint image B based on the image data of each rectangular template area Ai defined on the registered fingerprint image A and the detection of the area Bi are performed for each rectangular template area Ai. (Steps S3 to S5). When it is determined that each area Bi on the verification fingerprint image B having the maximum correlation coefficient has been detected in all the rectangular template areas Ai, the processing in the verification processing unit 23 is performed. As shown in FIG. 6, in the collation determination processing unit 27, the plurality of rectangular template areas ｛A defined in the registered fingerprint image A_i| I = 1, 2,..., N} (the mutual positional relationship) and the plurality of regions {B detected from the collation fingerprint image B_iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (steps S5 → S6).
[0062]
That is, in the collation determination processing (step S6), coordinates A1 specifying the positions of the rectangular template area A1 arranged on the registered fingerprint image A and the maximum correlation rectangular area B1 detected on the collation fingerprint image B on each image. Using (X0, Y0) and B1 (XD0, YD0) as reference coordinates, each of the rectangular template areas A2 to A5 arranged in the registered fingerprint image A and each of the maximum correlated rectangular areas B2 to B5 detected on the verification fingerprint image B For coordinates A2 (X1, Y1) to A5 (X4, Y4) and B2 (XD1, YD1) to B5 (XD4, YD4) for specifying the position on each image, A1 and A2, A3, A4, A5 .DELTA.i (i = 1, 2, 3, 4), which is the difference between the relative distances B1 and B2, B3, B4, and B5, is calculated based on the following equation.
[0063]
(Equation 1)

[0064]
Then, it is determined whether or not all the calculated Δi fall within a predetermined value. If all the Δi fall within the predetermined value, it is determined that the registered fingerprint A matches the collation fingerprint B. On the other hand, if not, the registered fingerprint A is determined. It is determined that it does not match the collation fingerprint B (step S7). Note that the predetermined value used here is, for example, a value that actually calculates Δi from fingerprint image data obtained from a plurality of persons and that obtains a desired matching accuracy based on the distribution of the calculation result.
[0065]
Here, when it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the collated fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). Are displayed as the same fingerprint (step S7 → end).
[0066]
On the other hand, if it is not determined that the registered fingerprint A and the collation fingerprint B are the same fingerprint, the template arrangement processing unit 25 subsequently performs the collation fingerprint after θm rotation stored in the collation fingerprint rotation image memory 13c. As shown in FIG. 8B, a plurality of rectangular template areas b_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged, and as shown in FIG. 9A, the coordinate data indicating the position of each template area bi is collated. The data is stored in the fingerprint position memory 13e (steps S7 → S8).
[0067]
Then, in the maximum correlation area detection processing unit 26 in the matching processing unit 23, the partial image corresponding to each rectangular template area bi defined and arranged with respect to the matching fingerprint image B after the rotation of θm is stored in the registered fingerprint data storage unit. 22, all the pixel data in the template bi and the corresponding registered fingerprint image A in the registered fingerprint image A are sequentially scanned while raster-scanning one pixel at a time in the horizontal and vertical directions on the registered fingerprint image A stored and registered in the storage device 12. A correlation coefficient using the pixel data is calculated. Then, an area ai in the registered fingerprint image A where the correlation coefficient becomes maximum is detected as shown in FIG. 8A, and as shown in FIG. 9B, the detection of each maximum correlation area ai is performed. The coordinate data indicating the position is stored in the registered fingerprint position memory 13d (steps S9, S10). The calculation of the correlation coefficient will be described later.
[0068]
The calculation of the maximum correlation coefficient on the registered fingerprint image A based on the image data of each rectangular template area bi defined on the collation fingerprint image B after the rotation of θm and the detection of the area ai are performed in each rectangular template area. The processing is sequentially performed for each bi (steps S9 to S11). When it is determined that each area ai on the registered fingerprint image A having the maximum correlation coefficient has been detected in all the rectangular template areas bi, the matching processing unit 23 As shown in FIG. 8, the plurality of rectangular template areas {b} defined in the rotated_i| I = 1, 2,..., N} and the plurality of areas {a detected from the registered fingerprint image A_iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (step S11 → S12).
[0069]
That is, in the collation determination processing (step S12), the positions of the rectangular template area b1 arranged on the collated fingerprint image B after the rotation of θm and the maximum correlation rectangular area a1 detected on the registered fingerprint image A are specified on each image. Coordinates b1 (X0, Y0), a1 (XD0, YD0), rectangular template areas b2 to b5 arranged in the verification fingerprint image B after rotation θm, and maximum correlation rectangular areas detected on the registered fingerprint image A The coordinates b2 (X1, Y1) to b5 (X4, Y4) and a2 (XD1, YD1) to a5 (XD4, YD4) specifying the position on each image of each of a2 to a5 are used, and b1 and b2, b3 are used. Δi (i = 1, 2, 3, 4), which is the difference between each relative distance to b4 and b5 and each relative distance to a1 and a2, a3, a4, and a5, is calculated based on the above equation.
[0070]
Then, it is determined whether or not all the calculated Δi fall within a predetermined value. If all the Δi fall within the predetermined value, it is determined that the registered fingerprint A matches the collation fingerprint B. On the other hand, if not, the registered fingerprint A is determined. It is determined that it does not match the collation fingerprint B (step S13).
[0071]
Here, when it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the collated fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). Are displayed as the same fingerprint (step S13 → end).
[0072]
On the other hand, if it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, all the rotational deviation correction angles θm (= 0 °, 1 °, −1 °, Each rotated verification fingerprint image B that has been rotationally converted by 2 °, −2 °, 3 °, −3 °,..., K °, −k ° [m = 1 to k × 2 + 1, k ≧ 1]. Then, it is determined whether or not the matching process with the registered fingerprint image A according to steps S1 to S13 has been performed (step S13 → S14).
[0073]
If it is determined that a series of matching processes with the registered fingerprint image A has not been performed for each rotated matching fingerprint image B that has been rotated and converted by all of the specified rotational deviation correction angles θm, Returning to the processing from step S1, the collation fingerprint image is subjected to the rotation conversion processing by the next rotational deviation correction angle θm (= m + 1), and the rotated collation fingerprint image B is converted into a template on the registered fingerprint image A in the same manner as described above. The collation processing when the area Ai is defined and arranged and the collation processing when the template area bi is defined and arranged on the rotated collation fingerprint image B are repeatedly performed (step S14 → S1 to S13).
[0074]
If it is determined that a series of collation processing with the registered fingerprint image A has been performed for each rotated collation fingerprint image B that has been rotationally converted by all the rotational deviation correction angles θm, the series of collation processing is performed. The process is ended, and the matching result display unit 28 (15) displays that the registered fingerprint A and the matching fingerprint B are not the same fingerprint (step S14 → end).
[0075]
Next, calculation of the correlation coefficient used in steps S3 to S5 and steps S9 to S11 of the fingerprint collation processing shown in FIGS. 4 and 5 will be described. Here, a description will be given as calculation of a correlation coefficient between the rectangular area A and the rectangular area B.
[0076]
First, pixels included in each of the rectangular area A and the rectangular area B are assumed to be A (i, j) and B (m, n), respectively. However, the total number of pixels included in each of the rectangular area A and the rectangular area B is made equal. Further, the signal intensities, which are multi-tone values indicating the shading of these pixels, are represented by_{ij ,}Y_mnAnd
[0077]
When these signal strengths are generalized and expressed as Zpq, the following equation is defined.
[0078]
(Equation 2)

[0079]
In the above equation, N indicates the total number of pixels included in the rectangular area. In the above equation, Σ indicates the sum of all the pixels included in the rectangular area. That is, the above expression indicates the average value of the signal intensity for the pixels included in the rectangular area.
[0080]
Next, the following equation is further defined.
[0081]
(Equation 3)

[0082]
The above equation shows the mean square value of the signal intensity for the pixels included in the rectangular area.
[0083]
Here, the correlation coefficient C between the rectangular area A and the rectangular area B_ABCan be calculated by the following equation expressed using the definition of the above equation.
[0084]
(Equation 4)

[0085]
The correlation coefficient between the regions is calculated using the above equation.
[0086]
Therefore, according to the fingerprint data matching device of the first embodiment having the above-described configuration, the rotation conversion process for performing the rotational deviation correction in consideration of the rotational center and the rotational deviation of the rotational deviation generated when the fingerprint image is collected. In many cases, the rotation center does not coincide with the specified rotation center, and the rotational displacement correction angle at which the collation match is obtained when the template is arranged on the registered fingerprint image A and the collation processing is performed, and the collation fingerprint image Focusing on the fact that the absolute value of the rotation deviation correction angle at which the collation match was obtained when the template was placed on B and the collation processing was performed is not equal, and each rotation including the rotational deviation correction angle θ1 = 0 ° was performed. Deviation correction angle θm (= 0 °, 1 °, −1 °, 2 °, −2 °, 3 °, −3 °,..., K °, −k ° [m = 1 to k × 2 + 1, k ≧ 1 Of the collation fingerprint image B rotated and converted by In this case, two types of matching processes, that is, a matching process when the template region Ai is arranged in the registered fingerprint image A and a matching process when the template region bi is arranged in the rotated fingerprint image B are performed alternately. Therefore, when performing fingerprint collation of the same person / same finger with rotational deviation, the possibility of obtaining a collation match at a smaller rotational deviation correction angle θm increases, and the amount of computation processing is reduced. The collation time can be reduced.
[0087]
In the first embodiment, only the verification fingerprint image B is subjected to the rotation conversion process for correcting the rotational deviation, and the registered fingerprint image A is rotationally converted by the rotational deviation correction angle θm. Two types of collation processing are performed alternately: a case where a template Ai is defined and arranged above and a case where a template bi is defined and arranged on the rotationally transformed collation fingerprint image B. As described in the second embodiment, a collation process performed by defining and arranging a template Ai on a non-rotated registered fingerprint image A for a collation fingerprint image B that is sequentially rotated and converted at a rotational misregistration correction angle θm; A configuration in which, for a registered fingerprint image A that is sequentially rotated and converted by θm, a matching process performed by defining and arranging a template bi on a non-rotating matching fingerprint image B is performed alternately. And, in the same manner as described above, it may be more collation coincidence in a small rotational shift correction angle θm is obtained.
[0088]
(2nd Embodiment)
FIG. 10 is a block diagram showing a configuration of an operation function of the fingerprint data matching device according to the second embodiment of the present invention.
[0089]
In the fingerprint matching operation function of the second embodiment, the rotation conversion processing unit 24 rotates the matching fingerprint image B given from the fingerprint data input unit 21 and the registered fingerprint image A given from the registered fingerprint data storage unit 22. When the collation fingerprint image B is rotated and converted alternately by the shift correction angle θm, the template arrangement processing unit 25 compares the registered fingerprint image A provided from the registered fingerprint data storage unit 22 with the template area Ai. When the registered fingerprint image A is rotationally transformed, a template area bi is defined and defined for the collation fingerprint image B provided from the fingerprint data input unit 21.
[0090]
The maximum correlation area detection processing unit 26 and the collation determination processing unit 27 define the collation fingerprint image B, which is sequentially rotated and converted by the rotation conversion processing unit 24, on the registered fingerprint image A by the template arrangement processing unit 25. A process of detecting the maximum correlation area Bi corresponding to the arranged template area Ai and collating the same, and comparing the registered fingerprint image A sequentially rotated and converted by the rotation conversion processing section 24 with the template arrangement processing section 25 The detection of the maximum correlation area ai corresponding to the template area bi defined and arranged on the fingerprint image B and the processing of collating the same are alternately performed.
[0091]
Next, a description will be given of the fingerprint matching function of the second embodiment in the fingerprint data matching device having the above-described configuration.
[0092]
FIG. 11 is a flowchart showing a fingerprint matching process (1) of the fingerprint data matching device according to the second embodiment.
[0093]
FIG. 12 is a flowchart showing a fingerprint matching process (part 2) of the second embodiment of the fingerprint data matching device.
[0094]
First, the entirety of the collation fingerprint image B, which is a multi-tone image collected and stored by the fingerprint data input unit 21, is subjected to the rotational displacement correction angle θm (= 0 °, 1 °, −1 °, 2 , -2 °, 3 °, -3 °,..., K °, -k ° [m = 1 to k × 2 + 1, integer of k ≧ 1]) (Step S1). However, when θ1 = 0, actual rotation conversion is not performed.
[0095]
Then, in the template arrangement processing unit 25, a plurality of rectangular template areas A_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged (step S2).
[0096]
Then, in the maximum correlation area detection processing section 26 in the matching processing section 23, the partial image corresponding to each rectangular template area Ai defined and arranged with respect to the registered fingerprint image A is subjected to the θm rotation conversion in the rotation conversion processing section 24. The relative relationship using all the pixel data in the template Ai and the corresponding pixel data in the collation fingerprint image B sequentially while raster-scanning one pixel at a time in the horizontal and vertical directions on the collated fingerprint image B The number is calculated. Then, an area Bi in the collation fingerprint image B having the maximum correlation coefficient is detected (steps S3 and S4).
[0097]
The calculation of the maximum correlation coefficient on the post-rotation collation fingerprint image B based on the image data of each rectangular template area Ai defined on the non-rotated registered fingerprint image A and the detection of the area Bi are performed by each rectangular template. The processing is sequentially performed for each area Ai (steps S3 to S5). If it is determined that each area Bi on the verification fingerprint image B having the maximum correlation coefficient has been detected in all the rectangular template areas Ai, the verification processing unit 23, the plurality of rectangular template areas {A} defined in the registered fingerprint image A._i| I = 1, 2,..., N} (the mutual positional relationship) and the plurality of regions {B detected from the collation fingerprint image B_iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (step S5 → S6).
[0098]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). The same fingerprint is displayed (step S7 → end).
[0099]
On the other hand, if it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, subsequently, the entire registered fingerprint image A stored in the registered fingerprint data storage unit 22 is processed by the rotation conversion processing unit 24. Rotational deviation correction angle θm (= 0 °, 1 °, −1 °, 2 °, −2 °, 3 °, −3 °,..., K °, −k ° [m = 1 to k × 2 + 1, k ≧ (An integer of 1)] (step S7 → S8a). However, when θ1 = 0, actual rotation conversion is not performed.
[0100]
Then, in the template arrangement processing unit 25, a plurality of rectangular template regions b_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged (step S8).
[0101]
Then, in the maximum correlation area detection processing unit 26 in the matching processing unit 23, the partial image corresponding to each rectangular template area bi defined and arranged for the non-rotated matching fingerprint image B is converted by the rotation conversion processing unit 24. While raster-scanning one pixel at a time in the horizontal and vertical directions on the registered fingerprint image A that has undergone the θm rotation conversion, all the pixel data in the template bi and the corresponding pixel data in the registered fingerprint image A are used sequentially. The calculated correlation coefficient is calculated. Then, the area ai in the registered fingerprint image A where the correlation coefficient becomes maximum is detected (steps S9 and S10).
[0102]
The calculation of the maximum correlation coefficient on the registered fingerprint image A after rotation based on the image data of each rectangular template area bi defined on the non-rotated verification fingerprint image B and the detection of the area ai are performed by each rectangular template. It is sequentially performed for each area bi (steps S9 to S11), and when it is determined that each area ai on the registered fingerprint image A having the maximum correlation coefficient has been detected in all the rectangular template areas bi, the matching processing unit 23, the plurality of rectangular template regions {b} defined in the non-rotated verification fingerprint image B._i| I = 1, 2,..., N} and the plurality of regions {a detected from the registered fingerprint image A after rotation._iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (step S11 → S12).
[0103]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). It is displayed that the fingerprints are the same (step S13 → end).
[0104]
On the other hand, if it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, all the rotational deviation correction angles θm (= 0 °, 1 °, −1 °, Each rotated verification fingerprint image B that has been rotationally converted by 2 °, −2 °, 3 °, −3 °,..., K °, −k ° [m = 1 to k × 2 + 1, k ≧ 1]. For each of the registered fingerprint images A after rotation, it is determined whether the matching process with the non-rotated registered fingerprint image A and the matching process with the non-rotated registered fingerprint image B according to steps S1 to S13 have been performed ( Step S13 → S14).
[0105]
Then, the matching process is performed on the non-rotated registered fingerprint image A with respect to the non-rotated registered fingerprint image A for each rotated collated fingerprint image B that has been rotationally converted by all of the specified rotational deviation correction angles θm. If it is determined that the collation process with the non-rotated collation fingerprint image B has not been performed, the process returns to the process from step S1, and the collation subjected to rotation conversion by the next rotational deviation correction angle θm (= m + 1). The collation processing when the template area Ai is defined and arranged on the non-rotated registered fingerprint image A for the fingerprint image B, and the non-rotated registered fingerprint image A for the fingerprint image A that has been rotationally transformed by the next rotational deviation correction angle θm (= m + 1). The collation processing when the template area bi is defined and arranged on the rotated collation fingerprint image B is repeatedly performed (step S14 → S1 to S13).
[0106]
In addition, the collation processing is performed on the collated fingerprint image B after each rotation, which has been rotationally converted by all rotational misalignment correction angles θm, with the non-rotated registered fingerprint image A, and the non-rotated If it is determined that the matching process with the matching fingerprint image B has been performed, the series of matching processes is terminated, and the matching result display unit 28 (15) determines that the registered fingerprint A and the matching fingerprint B are not the same fingerprint. Is displayed (step S14 → end).
[0107]
Therefore, even when the fingerprint data collating apparatus of the second embodiment performs fingerprint collation of the same person / same finger having rotational deviation, there is a high possibility that collation matching can be obtained at a smaller rotational deviation correction angle θm. , The amount of computation can be reduced and the matching time can be reduced.
[0108]
In the second embodiment, when the collation fingerprint image B is rotationally converted, the template area Ai is defined and arranged on the non-rotated registration fingerprint image A, and the collation processing with the rotated collation fingerprint image B is performed. At the same time, when the registered fingerprint image A is rotationally converted, the template bi is defined and arranged on the non-rotated collation fingerprint image B and the collation processing with the rotated registered fingerprint image A is performed. As described in the embodiment, the collation fingerprint image B and the registered fingerprint image A are alternately rotated and converted, and template regions bi and Ai are defined and defined for the rotated and converted collation fingerprint image B and the registered fingerprint image A. The matching process with the non-rotating registered fingerprint image A and the matching process with the non-rotating matching fingerprint image B are alternately performed. Collation coincidence is may be obtained.
[0109]
(Third embodiment)
FIG. 13 is a block diagram showing a configuration of an operation function associated with execution of the fingerprint matching processing program of the third embodiment of the fingerprint data matching device.
[0110]
In the fingerprint matching operation function of the third embodiment, the rotation conversion processing unit 24 rotates the matching fingerprint image B given from the fingerprint data input unit 21 and the registered fingerprint image A given from the registered fingerprint data storage unit 22 by rotation. As the rotational conversion is alternately performed by the shift correction angle θm, template regions bi and Ai are defined and arranged for the rotationally converted verification fingerprint image B and registered fingerprint image A.
[0111]
Then, the maximum correlation area detection processing unit 26 and the collation determination processing unit 27 compare the non-rotated registered fingerprint image A given from the registered fingerprint data storage unit 22 with the template fingerprint processing unit 25 after each rotation. B, a process for detecting the maximum correlation area ai corresponding to the template area bi defined and arranged on B, and a processing for detecting a non-rotated collated fingerprint image B provided from the fingerprint data input section 21; At 25, the processing of detecting the maximum correlation area Bi corresponding to the template area Ai defined and arranged on the registered fingerprint image A after each rotation and performing the matching process are performed alternately.
[0112]
Therefore, even with the fingerprint data collation apparatus of the third embodiment, when performing fingerprint collation of the same person / same finger with rotational deviation, there is a high possibility that collation coincidence is obtained at a smaller rotational deviation correction angle θm. , The amount of computation can be reduced and the matching time can be reduced.
[0113]
Further, in the second embodiment, when the collation fingerprint image B is rotationally transformed, the template area Ai is defined and arranged on the non-rotated registered fingerprint image A, and the collation with the rotated collation fingerprint image B is performed. And a process of defining and arranging a template area bi on a non-rotated collation fingerprint image B when the registered fingerprint image A is rotationally transformed and collating with the registered fingerprint image A after the rotation. In the state where the judgment cannot be obtained, the rotation deviation correction angle θm is sequentially updated and alternately repeated. However, as described in the following fourth embodiment, the rotational deviation correction angle θm (θ1 = 0 °), the maximum correlation coefficient MAXCC calculated in the collation processing on the collation fingerprint image B when the template area Ai is defined and arranged on the registration fingerprint image A, and the template on the collation fingerprint image B A maximum correlation coefficient maxcc to be calculated by the matching process with respect to the registered fingerprint image A in the case of defining arranged over preparative region bi to magnitude comparison. If the maximum correlation coefficient MAXCC when the template area Ai is defined and arranged on the registered fingerprint image A is larger, the registered fingerprint image A is sequentially rotated and converted by the rotation deviation correction angle θm. When the matching process performed by defining the template region Ai is selected and fingerprint matching is performed, and the maximum correlation coefficient maxcc when the template region bi is defined and placed on the matching fingerprint image B is larger, With the configuration in which the matching process is performed by defining the template area bi on the matching fingerprint image B while sequentially rotating and converting the registered fingerprint image A by the rotation shift correction angle θm, the fingerprint matching is performed, thereby further reducing the number of operations. The matching time may be shortened by the processing amount.
[0114]
(Fourth embodiment)
FIG. 14 is a block diagram showing a configuration of an operation function associated with execution of a fingerprint collation processing program of the fourth embodiment of the fingerprint data collation apparatus.
[0115]
In the fingerprint collation operation function of the fourth embodiment, when the collation fingerprint image B given from the fingerprint data input unit 21 is rotationally converted by the rotational conversion processing unit 24 at the rotational deviation correction angle θm, the registered fingerprint data storage unit 22 An area where a template area Ai is defined and arranged on a given registered fingerprint image A, and a maximum correlation coefficient is calculated on the rotated collation fingerprint image B by the maximum correlation area detection processing unit 26 and the collation determination processing unit 27. When the registered fingerprint image A provided from the registered fingerprint data storage unit 22 is rotationally converted at the rotational shift correction angle θm in the rotational conversion processing unit 24, the fingerprint data is input from the fingerprint data input unit 21. The template region bi is defined and arranged on the collation fingerprint image B to be detected, and the rotation is performed by the maximum correlation region detection processing unit 26 and the collation determination processing unit 27. It is assumed that a region where the maximum correlation coefficient is calculated is detected on the registered fingerprint image A and the fingerprint matching is performed. In this case, when the rotation deviation correction angle θm in the rotation conversion processing unit 24 is θ1 = 0 ° accompanying the initial setting, the template area A1 is registered on the registered fingerprint image A calculated in the maximum correlation area detection processing unit 26. The maximum correlation coefficient MAXCC for the collated fingerprint image B in the case where it is arranged and the maximum correlation coefficient maxcc for the registered fingerprint image A in the case where the template area b1 is arranged on the collated fingerprint image B are determined by a template arrangement fingerprint data selection processing unit. 29, the template arrangement fingerprint data selection processing unit 29 determines that the maximum correlation coefficient MAXCC when the template area A1 is defined and arranged on the registered fingerprint image A is larger than the rotation misalignment correction angle. In the processing after θ2 = 1 ° in which θm is updated, the collation obtained from the fingerprint data input unit 21 The fingerprint image B is fixedly provided to the rotation conversion processing unit 24, and the registered fingerprint image A obtained from the registered fingerprint data storage unit 22 is selected and determined so as to be fixedly provided to the template arrangement processing unit 25. Is fixed to the process of performing the collation with the collation fingerprint image B which is sequentially rotated and converted by defining and arranging the template area Ai. If it is determined that the maximum correlation coefficient maxcc in the case where the template area b1 is defined and arranged on the collation fingerprint image B is larger than the maximum correlation coefficient maxcc, the registration is performed in the processing after θ2 = 1 ° in which the rotational deviation correction angle θm is updated. The registered fingerprint image A obtained from the fingerprint data storage unit 22 is fixedly provided to the rotation conversion processing unit 24, and the collation fingerprint image B obtained from the fingerprint data input unit 21 is fixedly provided to the template arrangement processing unit 25. The selection is determined, the template area bi is defined and arranged in the collation fingerprint image B, and the process is collated with the registered fingerprint image A that is sequentially rotated and converted.
[0116]
Next, a description will be given of a fingerprint matching function according to the fourth embodiment in the fingerprint data matching device having the above-described configuration.
[0117]
FIG. 15 is a flowchart showing a fingerprint matching process (part 1) of the fourth embodiment of the fingerprint data matching device.
[0118]
FIG. 16 is a flowchart showing a fingerprint collation process (part 2) of the fourth embodiment of the fingerprint data collation device.
[0119]
First, the entirety of the collation fingerprint image B, which is a multi-tone image collected and stored by the fingerprint data input unit 21, is subjected to the rotational displacement correction angle θm (= 0 °, 1 °, −1 °, 2 , -2 °, 3 °, -3 °,..., K °, -k ° [m = 1 to k × 2 + 1, integer of k ≧ 1]) (Step S1).
[0120]
However, the actual rotation conversion is not performed only when the rotational deviation correction angle θm is θ1 = 0 ° by the initial setting, that is, only at the time of the first rotation conversion processing of the collation fingerprint image B in step S1.
[0121]
Then, in the template arrangement processing unit 25, a plurality of rectangular template areas A_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged (step S2).
[0122]
Then, in the maximum correlation area detection processing section 26 in the matching processing section 23, the partial image corresponding to each rectangular template area Ai defined and arranged with respect to the registered fingerprint image A is subjected to the θm rotation conversion in the rotation conversion processing section 24. All the pixel data in the template Ai and the corresponding collation are sequentially scanned while raster-scanning one pixel at a time in the horizontal and vertical directions on the collated fingerprint image B (substantially not rotated when θ1 = 0 °). A correlation coefficient using the pixel data in the fingerprint image B is calculated. Then, an area Bi in the collation fingerprint image B having the maximum correlation coefficient is detected (steps S3 and S4).
[0123]
Here, for the first rectangular template area A1 defined on the registered fingerprint image A, the maximum correlation coefficient value calculated when detecting the area B1 from the collation fingerprint image B is MAXCC, and the template arrangement fingerprint data selection processing unit 29.
[0124]
The calculation of the maximum correlation coefficient and the detection of the area Bi on the post-rotation collation fingerprint image B based on the image data of each rectangular template area Ai defined on the non-rotated registered fingerprint image A The process is sequentially performed for each template region Ai (steps S3 to S5). When it is determined that each region Bi on the verification fingerprint image B having the maximum correlation coefficient has been detected in all the rectangular template regions Ai, the verification process is performed. In the collation determination processing unit 27 in the unit 23, the plurality of rectangular template areas {A} defined in the registered fingerprint image A_i| I = 1, 2,..., N} (the mutual positional relationship) and the plurality of regions {B detected from the collation fingerprint image B_iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (step S5 → S6).
[0125]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). The same fingerprint is displayed (step S7 → end).
[0126]
On the other hand, when it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, the rotation deviation correction angle θm currently set in the rotation conversion processing unit 24 is θ1 = 0 ° accompanying the initial setting. Is determined (step S7 → S7a).
[0127]
Here, when it is determined that the rotation deviation correction angle θm currently set in the rotation conversion processing unit 24 is θ1 = 0 ° accompanying the initial setting, the rotation deviation correction angle θm is subsequently stored in the registered fingerprint data storage unit 22. The entire registered fingerprint image A is subjected to the rotation shift correction angle θm (= 0 °, 1 °, −1 °, 2 °, −2 °, 3 °, -3 °,. The rotation conversion is performed by k ° and −k ° (m = 1 to k × 2 + 1, k ≧ 1 integer)) (steps S7a → S8a).
[0128]
However, only when the rotational deviation correction angle θm is θ1 = 0 ° by the initial setting, that is, only at the time of the first rotational conversion processing of the registered fingerprint image A in step S8a, the actual rotational conversion is not performed.
[0129]
Then, in the template arrangement processing unit 25, a plurality of rectangular template regions b_i[I = 1 to N: an integer of N ≧ 2 (N = 5 in this case)] is defined and arranged (step S8).
[0130]
Then, in the maximum correlation area detection processing unit 26 in the matching processing unit 23, the partial image corresponding to each rectangular template area bi defined and arranged for the non-rotated matching fingerprint image B is converted by the rotation conversion processing unit 24. All the pixel data in the template bi are sequentially scanned while raster-scanning one pixel at a time in the horizontal and vertical directions on the registered fingerprint image A that has been rotated by θm (substantially not rotated when θ1 = 0 °). A correlation coefficient using the corresponding pixel data in the registered fingerprint image A is calculated. Then, the area ai in the registered fingerprint image A where the correlation coefficient becomes maximum is detected (steps S9 and S10).
[0131]
Here, the maximum correlation coefficient value calculated when detecting the area a1 from the registered fingerprint image A for the first rectangular template area b1 defined on the collation fingerprint image B is set to maxcc, and the template arrangement fingerprint data selection processing unit 29.
[0132]
The calculation of the maximum correlation coefficient on the registered fingerprint image A after the rotation process based on the image data of each rectangular template region bi defined on the non-rotated collation fingerprint image B and the detection of the region ai are performed by each rectangle. The process is sequentially performed for each template region bi (steps S9 to S11). If it is determined that each region ai on the registered fingerprint image A having the maximum correlation coefficient has been detected in all the rectangular template regions bi, the collation processing is performed. In the collation determination processing unit 27 in the unit 23, the plurality of rectangular template areas {b_i| I = 1, 2,..., N} and the plurality of areas {a detected from the registered fingerprint image A_iThe distribution (mutual positional relationship) of | i = 1, 2,..., N} is compared, and the identity of the registered fingerprint A and the verification fingerprint B is evaluated (step S11 → S12).
[0133]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). It is displayed that the fingerprints are the same (step S13 → end).
[0134]
On the other hand, when it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, the rotation deviation correction angle θm currently set in the rotation conversion processing unit 24 is θ1 = 0 ° accompanying the initial setting. Is determined (step S13 → S13a).
[0135]
Here, when it is determined that the rotation deviation correction angle θm currently set in the rotation conversion processing unit 24 is θ1 = 0 ° accompanying the initial setting, the template arrangement fingerprint data selection processing unit 29 determines the maximum value. The maximum correlation coefficient value MAXCC with the collation fingerprint image B for the first rectangular template area A1 arranged on the registered fingerprint image A provided from the correlation area detection processing unit 26 and the first correlation coefficient value MAXCC arranged on the collation fingerprint image B A comparison is made between the rectangular template area b1 and the maximum correlation coefficient value maxcc with the registered fingerprint image A (step S13a → S15).
[0136]
Then, the maximum correlation coefficient value MAXCC of the first rectangular template area A1 arranged on the registered fingerprint image A with the collation fingerprint image B is registered in the first rectangular template area b1 arranged on the collated fingerprint image B. If it is determined that the value is larger than (or greater than) the maximum correlation coefficient value maxcc with the fingerprint image A, the rotation conversion processing unit 24 sequentially updates and sets the rotational deviation correction angle θm. In step S1 to step S7, the template area Ai is defined and arranged in the registered fingerprint image A, and the collation fingerprint image B is sequentially rotated and converted to perform fingerprint collation (steps S15 to S1 to S7). .
[0137]
In step S7, when it is not determined that the registered fingerprint A and the verification fingerprint B are the same fingerprint, all the rotational deviation correction angles θm (= 0 °, 1 °, -1 °, 2 °, -2 °, 3 °, -3 °,..., K °, -k ° [m = 1 to k × 2 + 1, k ≧ 1 integer] after each rotation It is determined whether or not the collation processing in which the template area Ai is arranged on the non-rotated registered fingerprint image A has been performed for each of the collation fingerprint images B (step S7 → S14a).
[0138]
Then, it is determined that the collation processing in which the template area Ai is arranged on the non-rotated registered fingerprint image A has not been performed for each rotated collation fingerprint image B that has been rotationally converted by all the rotation deviation correction angles θm specified above. In this case, the flow returns to the processing from step S1 again, and the template area Ai is placed on the non-rotated registered fingerprint image A of the collation fingerprint image B that has been rotationally converted by the updated rotational deviation correction angle θm (= m + 1). The collation processing defined and arranged is repeatedly performed (step S14a → S1 to S7).
[0139]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). The same fingerprint is displayed (step S7 → end).
[0140]
In addition, when it is determined that the matching process of the non-rotated registered fingerprint image A has been performed for the rotated matching fingerprint image B that has been rotated and converted by all the rotational deviation correction angles θm, the series of matching processes is performed. The process is terminated, and the matching result display unit 28 (15) displays that the registered fingerprint A and the matching fingerprint B are not the same fingerprint (step S14a → end).
[0141]
On the other hand, in the step S15, the maximum correlation coefficient value maxcc with the registered fingerprint image A for the first rectangular template area b1 arranged on the collation fingerprint image B is the first rectangle arranged on the registered fingerprint image A. If it is determined that the maximum correlation coefficient value MAXCC between the template area A1 and the collation fingerprint image B is larger than the maximum correlation coefficient value MAXCC, the rotation shift correction angle θm in the rotation conversion processing unit 24 is sequentially updated and set after θ2 = 1 °. In the processing, the fingerprint collation processing in steps S8a to S13 for defining and arranging the template area bi in the collation fingerprint image B and sequentially rotating and converting the registered fingerprint image A to perform fingerprint collation is performed (step S15 → S8a to S13) ).
[0142]
In this step S13, when it is not determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, all the rotational deviation correction angles θm (= 0 °, 1 °, -1 °, 2 °, -2 °, 3 °, -3 °,..., K °, -k ° [m = 1 to k × 2 + 1, k ≧ 1 integer] after each rotation For each of the registered fingerprint images A, it is determined whether or not the collation processing in which the template region bi is arranged on the non-rotated collation fingerprint image B has been performed (step S13 → S14b).
[0143]
Then, it is determined that the matching process of arranging the template area bi on the non-rotating matching fingerprint image B has not been performed for each rotated registered fingerprint image A that has been rotated and converted by all of the specified rotational deviation correction angles θm. If so, the process returns to step S8a again, and the template area bi is placed on the non-rotated verification fingerprint image B of the registered fingerprint image A that has been rotationally converted by the updated rotational deviation correction angle θm (= m + 1). The collation processing defined and arranged is repeatedly performed (step S14b → S8a to S13).
[0144]
When it is determined that the registered fingerprint A and the collated fingerprint B are the same fingerprint, the series of collation processing is terminated, and the registered fingerprint A and the collated fingerprint B are displayed by the collation result display unit 28 (15). It is displayed that the fingerprints are the same (step S13 → end).
[0145]
In addition, when it is determined that the matching process of the non-rotated collation fingerprint image B is performed for the registered fingerprint image A after each rotation that has been rotationally converted by all the rotational deviation correction angles θm, the series of collation is performed. The process is terminated, and the matching result display unit 28 (15) displays that the registered fingerprint A and the matching fingerprint B are not the same fingerprint (step S14b → end).
[0146]
Therefore, according to the fingerprint data matching device of the fourth embodiment, a fingerprint matching process for defining and arranging the template region Ai in the registered fingerprint image A and matching with the matching fingerprint image B which is sequentially rotated and converted, The fingerprint collation processing for defining and arranging the template area bi in the image B and collating with the registered fingerprint image A which is sequentially rotated and converted is fixed to the fingerprint collation processing which can obtain a large correlation coefficient value in each initial routine. In the same manner as in the first to third embodiments, in order to perform fingerprint matching of the same person / same finger having a rotational displacement, matching is performed at a smaller rotational displacement correction angle θm. Not only is the probability of being obtained higher, but also the collation time can be greatly reduced with a smaller amount of computation.
[0147]
In step S15 in the fingerprint collation processing of the fourth embodiment, the maximum correlation coefficient value MAXCC with the collation fingerprint image B for the first rectangular template area A1 arranged on the registered fingerprint image A, Is compared with the maximum correlation coefficient value maxcc with the registered fingerprint image A with respect to the first rectangular template area b1 arranged in the first rectangular template area b1, but instead of this, a plurality of rectangular templates arranged on the registered fingerprint image A are replaced. The average value of the maximum correlation coefficient values MAXCC with the matching fingerprint image B for each of the regions Ai, and the maximum correlation coefficient value with the registered fingerprint image A for each of the plurality of rectangular template regions bi arranged on the matching fingerprint image B A configuration may be employed in which the magnitude of the maxcc is compared with the average value.
[0148]
Further, in the fourth embodiment, when the collation fingerprint image B is rotationally converted, the template area Ai is defined and arranged on the non-rotated registered fingerprint image A, and the collation processing with the collation fingerprint image B after the rotation processing is performed. When the registered fingerprint image A is rotationally converted, the template bi is defined and arranged on the non-rotated collation fingerprint image B, and the collation processing with the registered fingerprint image A after the rotation processing is performed. As described in the embodiment, template regions bi and Ai are defined and arranged for a rotationally transformed verification fingerprint image B or registration fingerprint image A, and a verification process with a non-rotation registration fingerprint image A or a non-rotation verification fingerprint is performed. It may be configured to perform the collation processing with the image B.
[0149]
(Fifth embodiment)
FIG. 17 is a block diagram showing a configuration of an operation function associated with execution of the fingerprint matching processing program of the fifth embodiment of the fingerprint data matching apparatus.
[0150]
In the fingerprint collation operation function of the fifth embodiment, the registered fingerprint image A is rotationally converted by the rotation conversion processing unit 24 through the template arrangement fingerprint data selection processing unit 29, and the rotationally converted registered fingerprint image A is subjected to the template arrangement processing unit When the template area Ai is defined and arranged in step 25, the non-rotated verification fingerprint image B is directly supplied to the maximum correlation area detection processing section 26 through the template arrangement fingerprint data selection processing section 29, and the template arrangement fingerprint data selection processing is performed. When the collation fingerprint image B is rotated and converted by the rotation conversion processing unit 24 through the unit 29 and the template area bi is defined and arranged for the rotation-converted collation fingerprint image B by the template arrangement processing unit 25, the template arrangement fingerprint data The non-rotated registered fingerprint image A is directly subjected to the maximum correlation area detection processing through the selection processing unit 29. Give to part 26.
[0151]
Then, at the initially set rotational deviation correction angle θm (θ1 = 0 °), the maximum value calculated between the registered fingerprint image A and the non-rotated verification fingerprint image B by rotationally converting the registered fingerprint image A to define and arrange the template area Ai. Between the correlation coefficient value MAXCC and the initially set rotational deviation correction angle θm (θ1 = 0 °), the collation fingerprint image B is rotationally converted to define and arrange the template area bi to register the non-rotated registered fingerprint image A. The maximum correlation coefficient value maxcc calculated in step (1) is supplied from the maximum correlation area detection processing unit 26 to the template arrangement fingerprint data selection processing unit 29 to determine the magnitude, and the rotational deviation correction angle θm, which is sequentially updated, is θ2 = 1. In the subsequent processing, the fingerprint fingerprint processing is performed by performing a rotational transformation of the registered fingerprint image A, defining and arranging the template area Ai, and performing the rotation with the non-rotated verification fingerprint image B, or A structure for selecting determines a fingerprint collation processing performed with the registered fingerprint image A rotation transformation defined arranged to non-rotationally template regions bi collation fingerprint image B.
[0152]
Therefore, in the fingerprint data collation device of the fifth embodiment, as in the fourth embodiment, the fingerprint collation of the same person and the same finger having a rotational deviation is performed at a smaller rotational deviation correction angle θm. Not only is it more likely that a match will be obtained, but it will also be possible to significantly reduce the collation time with a smaller amount of computation.
[0153]
In the third embodiment and the fifth embodiment, in the rotation conversion processing of the registered fingerprint image A or the collation fingerprint image B on the side where the template region is defined and arranged, the entire fingerprint image A or B is rotationally converted. Although the configuration has been described, the configuration may be such that only a partial fingerprint image area that is a target of the definition arrangement of the template area of the fingerprint image A or B is rotationally converted. According to this, the amount of the image to be subjected to the rotation conversion processing itself can be reduced, and the rotation conversion processing time can be shortened to further reduce the fingerprint collation processing time.
[0154]
In the matching process between the registered fingerprint image A and the matching fingerprint image B in each of the above embodiments, each of the regions Ai, Bi, bi, and ai is a rectangular region. However, the shape of these regions is not limited to a rectangle and may be any shape. Is fine. It is desirable that the sizes and shapes of the regions Ai, Bi, bi, and ai are the same, but even if there is some difference, it is acceptable as long as the required fingerprint matching accuracy is obtained. Further, the positions where the template regions Ai and bi are defined and arranged may be arbitrary.
[0155]
In the collation determination, in addition to the method of evaluating Δi in the above-described processing, for example, the shape or the area of a figure formed with Ai or bi as a vertex and a figure formed with Bi or ai as a vertex is determined. It is also possible to adopt various methods such as a judgment based on a difference.
[0156]
The method described in each of the above embodiments, that is, the fingerprint collation processing in each of the first to fifth embodiments based on each operation function block diagram and each flowchart in FIGS. 3 to 5 and FIGS. Each method is a program that can be executed by a computer, such as a memory card (ROM card, RAM card, etc.), a magnetic disk (floppy disk, hard disk, etc.), an optical disk (CD-ROM, DVD, etc.), an external memory such as a semiconductor memory, etc. It can be stored in the storage medium 18a and distributed. Then, the computer of the fingerprint data collation device reads the program stored in the external storage medium 18a into the storage device 12, and the operation is controlled by the read program, thereby performing the fingerprint collation function described in the above embodiment. And can execute the same processing by the method described above.
[0157]
Further, data of a program for realizing each of the above methods can be transmitted on a network (public line) 20 in the form of a program code, and the transmission control unit 19 of the fingerprint data collation apparatus connected to the network 20 Thus, the above-described program data can be fetched to realize the above-described fingerprint collation function.
[0158]
【The invention's effect】
As mentioned above,According to the present invention, when the image data A and the image data B are collated, two types of collation processing of rotating the image and alternately arranging the templates on the image A and the image B are performed. Even if a match cannot be obtained by the matching process with the template placed on the, there is a possibility that the match will be obtained by the matching process with the template placed on the other,Image DayOfRotation processing timesNumberPicture while smallStatue AAnd B are more likely to be matched and the matching time can be reduced.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of an electronic circuit of a fingerprint data matching device according to an embodiment of an image data matching device of the present invention.
FIG. 2 is a diagram showing a data memory secured in a RAM of the fingerprint data matching device.
FIG. 3 is a block diagram showing a configuration of an operation function accompanying execution of a fingerprint matching processing program of the first embodiment of the fingerprint data matching apparatus.
FIG. 4 is a flowchart showing a fingerprint collation process (part 1) of the first embodiment of the fingerprint data collation device;
FIG. 5 is a flowchart showing a fingerprint matching process (part 2) of the first embodiment of the fingerprint data matching device.
FIG. 6 shows a comparison between the arrangement position distribution of the template area Ai on the registered fingerprint image A and the detection position distribution of the maximum correlation area Bi on the collation fingerprint image B in the fingerprint collation processing of the fingerprint data collation apparatus. FIG.
FIG. 7 is a diagram showing respective position data of a template position and a maximum correlation area detection position stored in a registered fingerprint position memory and a collated fingerprint position memory in a RAM in accordance with the fingerprint collation processing of the fingerprint data collation device.
FIG. 8 shows a comparison state between the arrangement position distribution of the template area bi on the collation fingerprint image B and the detection position distribution of the maximum correlation area ai on the registered fingerprint image A accompanying the fingerprint collation processing of the fingerprint data collation apparatus. FIG.
FIG. 9 is a view showing respective position data of a template position and a maximum correlation area detection position stored in a collation fingerprint position memory and a registered fingerprint position memory in a RAM in accordance with the fingerprint collation processing of the fingerprint data collation device.
FIG. 10 is a block diagram showing a configuration of an operation function accompanying execution of a fingerprint matching processing program of the second embodiment of the fingerprint data matching apparatus.
FIG. 11 is a flowchart showing a fingerprint matching process (1) of the fingerprint data matching device according to the second embodiment.
FIG. 12 is a flowchart showing a fingerprint collation process (part 2) of the second embodiment of the fingerprint data collation device.
FIG. 13 is a block diagram showing a configuration of an operation function associated with execution of a fingerprint matching processing program according to a third embodiment of the fingerprint data matching device.
FIG. 14 is a block diagram showing a configuration of an operation function accompanying execution of a fingerprint collation processing program according to a fourth embodiment of the fingerprint data collation device.
FIG. 15 is a flowchart showing a fingerprint collation process (part 1) of the fingerprint data collation device according to a fourth embodiment.
FIG. 16 is a flowchart showing a fingerprint matching process (part 2) of the fingerprint data matching device according to a fourth embodiment.
FIG. 17 is a block diagram showing a configuration of an operation function associated with execution of a fingerprint matching processing program according to a fifth embodiment of the fingerprint data matching apparatus.
FIG. 18 is a flowchart showing a procedure of fingerprint data matching in the fingerprint data matching device of the prior application.
FIG. 19 is a diagram showing a plurality of reference regions arranged in registered fingerprint data in accordance with the fingerprint data collation procedure of the prior application and a plurality of regions detected on the collation fingerprint data based on the reference regions.
FIG. 20 shows an example in which a plurality of reference areas arranged in registered fingerprint data in accordance with the fingerprint data collation procedure of the prior application and a plurality of inappropriate areas on the collation fingerprint data are detected based on the reference areas. FIG.
[Explanation of symbols]
11 ... control unit (CPU)
12 ... storage device
13… RAM
13a: Read image memory
13b ... collation fingerprint image memory
13c: Verification fingerprint rotation image memory
13d: Registered fingerprint position memory
13e: collation fingerprint position memory
13f: rotation angle memory
14 ... Image reading device
15 Display unit
16 Input unit
17… Bus
18 ... storage medium reading unit
18a: external storage medium
19: Transmission control unit
20… Network
20a: Program server (external computer terminal)
20b: external terminal storage device
21 ... fingerprint data input section
22 ... registered fingerprint data storage
23 ... collation processing unit
24 ... rotation conversion processing unit
25 ... template arrangement processing unit
26 ... maximum correlation area detection processing unit
27… collation judgment processing section
28… collation result display
29 ... template arrangement fingerprint data selection processing unit

Claims (9)

An image data collation device that collates image data A and image data B to determine the identity of the two,
Image rotating means for rotating by a predetermined angle with one of the image data A or the image data B,
First template arranging means for arranging a template defining a plurality of regions for the image data A for each rotation of the image data A or B by the image rotator;
First correlation area detection means for detecting, on the image data B, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data A by the first template arrangement means;
The mutual positional relationship between the plurality of regions arranged in the image data A by the first template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting unit are determined. First collation determination means for comparing and determining the identity of the image data A and the image data B;
A second template arranging unit for arranging a template defining a plurality of regions for the image data B for each rotation of the image data A or B by the image rotating unit;
A second correlation area detecting means for detecting, on the image data A, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data B by the second template arranging means;
The mutual positional relationship between the plurality of regions arranged in the image data B by the second template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the second correlation region detecting unit are determined. Second collation determination means for comparing and determining the identity between the image data A and the image data B;
An image data collation device comprising: An image data collation device that collates image data A and image data B to determine the identity of the two,
Image rotating means for alternately rotating the image data A and the image data B by a preset angle,
When the image data A is rotated by the image rotation means, a template defining a plurality of regions is arranged for the non-rotated image data B, and when the image data B is rotated, the template is added to the non-rotated image data A. A template arranging means for arranging a template defining a plurality of areas with respect to the
When a template is arranged in the image data B by the template arranging means, a first correlation for detecting, on the image data A, a plurality of correlation regions having a maximum correlation with each of the plurality of regions defined by the template. Area detection means;
A comparison is made between the mutual positional relationship between the plurality of regions arranged in the image data B by the template arranging means and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the first correlation region detecting means. First collation determination means for determining the identity between the image data A and the image data B;
When a template is arranged in the image data A by the template arranging means, a second correlation for detecting, on the image data B, a plurality of correlation regions having a maximum correlation with each of the plurality of regions defined by the template. Area detection means;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the second correlation region detecting unit. Second collation determination means for determining the identity between the image data A and the image data B;
An image data collation device comprising: An image data collation device that collates image data A and image data B to determine the identity of the two,
Image rotating means for alternately rotating image data A and image data B by a preset angle Steps and
When the image data A is rotated by the image rotating means, a template defining a plurality of regions is arranged for the image data A, and when the image data B is rotated, a plurality of templates are defined for the image data B. Template placement means for placing a template defining an area of
When a template is arranged on the image data A by the template arranging means, a plurality of correlation areas having the maximum correlation with each of the plurality of areas defined by the template are detected on the non-rotated image data B. 1 correlation area detecting means;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting unit. First collation determination means for determining the identity between the image data A and the image data B;
When a template is arranged in the image data B by the template arranging means, a plurality of correlation regions having the maximum correlation with each of the plurality of regions defined by the template are detected on the non-rotated image data A. 2 correlation area detecting means;
By comparing a mutual positional relationship between a plurality of regions arranged in the image data B by the template arranging unit and a mutual positional relationship between a plurality of correlation regions detected on the image data A by the second correlation region detecting unit. Second collation determination means for determining the identity between the image data A and the image data B;
An image data collation device comprising: An image data matching method for comparing image data A and image data B to determine the identity of the two,
An image rotation step for rotating by a predetermined angle with one of the image data A or the image data B,
A first template arranging step of arranging a template defining a plurality of regions with respect to the image data A for each rotation of the image data A or B in the image rotating step;
A first correlation area detection step of detecting, on the image data B, a plurality of correlation areas having the maximum correlation with each of the plurality of areas arranged in the image data A by the first template arrangement step;
The mutual positional relationship between the plurality of regions arranged in the image data A by the first template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting step are determined. A first collation determination step of comparing and determining the identity of the image data A and the image data B;
A second template arranging step of arranging a template defining a plurality of regions for the image data B for each rotation of the image data A or B in the image rotating step;
A second correlation area detection step of detecting, on the image data A, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data B by the second template arrangement step;
The mutual positional relationship between the plurality of regions arranged in the image data B by the second template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the second correlation region detecting step are determined. A second collation determining step of comparing and determining the identity of the image data A and the image data B;
An image data collating method comprising: An image data collation method for collating image data A and image data B to determine the identity of the two,
An image rotation step of alternately rotating the image data A and the image data B by a preset angle,
When the image data A is rotated by the image rotation step, a template defining a plurality of regions is arranged for the non-rotated image data B, and when the image data B is rotated, the template is added to the non-rotated image data A. Template to place a template that defines multiple areas Port placement step;
When a template is arranged in the image data B by the template arranging step, a first correlation for detecting, on the image data A, a plurality of correlation regions having a maximum correlation with each of the plurality of regions defined by the template. An area detection step;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data B by the template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the first correlation region detecting step. A first collation determination step of determining the identity between the image data A and the image data B;
A second correlation for detecting, on the image data B, a plurality of correlation regions having the maximum correlation with each of the plurality of regions defined by the template when the template is arranged in the image data A by the template arranging step; An area detection step;
Comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the second correlation region detecting step; A second collation determination step of determining the identity between the image data A and the image data B;
An image data collating method comprising: An image data collation method for collating image data A and image data B to determine the identity of the two,
An image rotation step of alternately rotating the image data A and the image data B by a preset angle,
When the image data A is rotated in the image rotation step, a template defining a plurality of regions is arranged for the image data A, and when the image data B is rotated, a plurality of templates are defined for the image data B. A template arrangement step of arranging a template defining an area of
When a template is arranged in the image data A by the template arranging step, a plurality of correlation regions having the maximum correlation with each of the plurality of regions defined by the template are detected on the non-rotated image data B. 1 correlation region detection step;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting step. A first collation determination step of determining the identity between the image data A and the image data B;
When a template is arranged in the image data B by the template arranging step, a plurality of correlation regions having the maximum correlation with each of the plurality of regions defined by the template are detected on the non-rotated image data A. 2 correlation area detection steps;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data B by the template arranging step and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the second correlation region detecting step, A second collation determination step of determining the identity between the image data A and the image data B;
An image data collating method comprising: A storage medium storing an image data collation processing program for controlling a computer to collate image data A and image data B to determine the identity of the two,
Said computer,
Image rotating means for rotating either the image data A or the image data B by a preset angle,
First template arranging means for arranging a template defining a plurality of regions for the image data A for each rotation of the image data A or B by the image rotator;
First correlation area detection means for detecting, on the image data B, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data A by the first template arrangement means;
The mutual positional relationship between the plurality of regions arranged in the image data A by the first template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting unit are determined. First collation determining means for comparing and determining the identity of the image data A and the image data B;
A second template arranging means for arranging a template defining a plurality of regions for the image data B for each rotation of the image data A or B by the image rotator;
A second correlation area detecting means for detecting, on the image data A, a plurality of correlation areas having a maximum correlation with each of the plurality of areas arranged in the image data B by the second template arranging means;
The mutual positional relationship between the plurality of regions arranged in the image data B by the second template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the second correlation region detecting unit are determined. Second collation determination means for comparing and determining the identity of the image data A and the image data B;
A storage medium storing a computer-readable image data collation processing program that is caused to function as a computer. A storage medium storing an image data collation processing program for controlling a computer to collate image data A and image data B to determine the identity of the two,
Said computer,
Image rotating means for alternately rotating the image data A and the image data B by a preset angle,
When the image data A is rotated by the image rotation means, a template defining a plurality of regions is arranged for the non-rotated image data B, and when the image data B is rotated, the template is added to the non-rotated image data A. A template arranging means for arranging a template defining a plurality of areas with respect to the
When a template is arranged in the image data B by the template arranging means, a first correlation for detecting, on the image data A, a plurality of correlation regions having a maximum correlation with each of the plurality of regions defined by the template. Area detection means,
A comparison is made between the mutual positional relationship between the plurality of regions arranged in the image data B by the template arranging means and the mutual positional relationship between the plurality of correlation regions detected on the image data A by the first correlation region detecting means. First collation determination means for determining the identity between the image data A and the image data B;
When a template is arranged in the image data A by the template arranging means, a second correlation for detecting, on the image data B, a plurality of correlation regions having a maximum correlation with each of the plurality of regions defined by the template. Area detection means,
By comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the second correlation region detecting unit. Second collation determination means for determining the identity between the image data A and the image data B;
A storage medium storing a computer-readable image data collation processing program that is caused to function as a computer. A storage medium storing an image data collation processing program for controlling a computer to collate image data A and image data B to determine the identity of the two,
Said computer,
Image rotating means for alternately rotating the image data A and the image data B by a preset angle,
When the image data A is rotated by the image rotating means, a template defining a plurality of regions is arranged for the image data A, and when the image data B is rotated, a plurality of templates are defined for the image data B. Template arranging means for arranging a template defining an area of
When a template is arranged in the image data A by the template arranging means, a plurality of phases having the maximum correlation with each of the plurality of regions defined by the template are set. First correlation area detection means for detecting the correlation area on the non-rotated image data B;
By comparing the mutual positional relationship between the plurality of regions arranged in the image data A by the template arranging unit and the mutual positional relationship between the plurality of correlation regions detected on the image data B by the first correlation region detecting unit. First collation determination means for determining the identity between the image data A and the image data B;
When a template is arranged in the image data B by the template arranging means, a plurality of correlation regions having the maximum correlation with each of the plurality of regions defined by the template are detected on the non-rotated image data A. 2. Correlation area detection means,
By comparing a mutual positional relationship between a plurality of regions arranged in the image data B by the template arranging unit and a mutual positional relationship between a plurality of correlation regions detected on the image data A by the second correlation region detecting unit. Second collation determination means for determining the identity between the image data A and the image data B;
A storage medium storing a computer-readable image data collation processing program that is caused to function as a computer.

Images