1281642 with the second image compressing data to output an image compressing information. In addition，the method provides a coding algorithmic process for the foregoing coding processes. And then， the method attains a better image effect by compressing the digital image that can transmit with a limited transmitting quantity. 七、 指定代表圖·· (一） 本案指定代表圖為：第七圖。 (二） 本代表圖之元件符號簡單說明： S71〜S77 :流程步驟。 八、 本案若有化學式時，請揭示最能顯示發明特徵的化學式： 九、 發明說明： 【發明所屬之技術領域】 本發明係揭露一種圖像壓縮/解壓縮方法及其圖像 編解碼系統，特別是關於圖像壓縮/解壓縮方法。 【先前技術】 目前可攜式通訊裝置已經成為隨身必備的通信產 品’同日守多媒體間訊服務（Multimedia Message Service， 4 ⑧ 1281642 MMS)和拍攝功能已成為可攜式通訊裝置的基本配置。 且可攜式通訊裝置的拍攝解析度從十萬像素一躍至百 萬像素’提南了拍攝圖片的晝質，若欲以刪傳送此高 f質，其他人分享’則受限於簡訊服務的流量。目 刖市場上所提供的多媒體簡訊服務流量限制在100K以 内（3G為300K)，而非動晝圖片則限制在了 5〇κ以内。 如此’ I法滿足百萬像素拍攝的圖片傳送需要I·以 上大小的流量需求，而使用者無法將所拍攝的高晝質圖 片分享給其他朋友。再者，目前可擴式通訊裝置大多支 援的圖片存儲格式有Mp、GIF、pNG和娜，其中當 要用於MMS k ’最佳的格式是壓縮率較高且圖像品質可 Z受的JPEG。但即使採用刑G格式保存一幅採用百 :::6.8Xl〇24解析度色彩較豐富的圖片，壓縮率採 等祕度也需要12GK左右的存儲空間，若加大壓 參 =或者減小圖片解析度以及分割圖片都將使圖片 硯員性和完整性得到較大損失。 明麥閱弟一圖’係為習知技藝之圖像壓縮技術之流 :圖。此流程由開始展開，首先提供—原始圖像⑴， 、、坐由數位採樣後進行數學轉換S12，例如町㈤咖工 =ln!transf则）轉換，接著，於數學轉換後的資料 仃，化處理S13 ’再將量化處理的結果進行編碼 後’於編碼後，齡ψ [51 /A i^r> 束整個流程。 輪出圖像I缩資料S15,結 1281642 加Μ近年來小波轉換的圖像壓縮技術是當前研究的一 ^^領域’它可以取得比傳統的娜更好的圖像廢 :::。請參閱第二圖’係為習知技藝之小波轉換圖像 ^技術之流程圖。此流程由開始展開，提供-圖像資 二/21 «I以塔式分解的方式對圖像資料進行小波轉換 ，接著’將轉換後的資料以EZW( Embedded Zero-tree avelet Encoding) ^ SPIHT ( Set Partitioning In erarchical Trees Coding)的方法進行量化 S23， 接著？字量化的結果於進行編碼後輸出圖像壓縮數據 S24 ’最後，結束整個流程。 口月茶閱第。圖’係為習知技藝之小波轉換圖像解壓 縮技術之流程圖。此流程由開始展開，提供一圖像壓縮 =據S3卜接著’針對圖像壓縮數據進行解碼，將 解碼後的資料執行逆量化S33，其後，再將逆量化的結 果進行小波轉換的逆轉換S34,然後，於逆轉換後輸出 圖像資料S35，則結束整個流程。 請參閱第四圖，係為習知技藝之小波轉換之示意 圖。以一 4x4像素的原始圖像41為例，其中，小波轉 換係對原始圖像41進行一水平分割42、一垂直分割43 兩個步驟。首先，進行水平分割42，此原始圖像41之 第一列像素標示為A、β、c及D，於水平分割42之第 -列係為A，、B，、C，及D，子圖，而Α，係為α+°β、β，係為 C+D、C係為Α-Β及D係為c-D ’同樣的第二列亦有Ε，、 FH Η，子圖’則冑人，、『^，子圖區分為低頻 1281642And the second, the method attains a better image effect by compressing the digital image that can transmit with a Limited transmitting quantity. VII. Designated representative map (1) The representative representative of the case is: the seventh picture. (2) A brief description of the component symbols of this representative figure: S71~S77: Process steps. 8. If there is a chemical formula in this case, please disclose the chemical formula that best shows the characteristics of the invention: IX. Description of the invention: [Technical field of the invention] The present invention discloses an image compression/decompression method and an image encoding and decoding system thereof. Especially with regard to image compression/decompression methods. [Prior Art] At present, the portable communication device has become a necessary communication product. The Multimedia Message Service (4 8 1281642 MMS) and the shooting function have become the basic configuration of the portable communication device. And the resolution of the portable communication device jumps from 100,000 pixels to megapixels. 'Tenan took the temperament of taking pictures. If you want to delete this high quality, others share 'it is limited by the newsletter service. flow. The multimedia newsletter service provided in the market is limited to 100K (3K is 300K), while the non-moving picture is limited to 5〇κ. In this way, the image transmission required to satisfy the megapixel shooting requires a flow demand of more than one size, and the user cannot share the photographed high quality picture to other friends. Furthermore, most of the current scalable communication devices support image storage formats such as Mp, GIF, pNG, and Na, which are used in MMS k. The best format is JPEG with high compression ratio and image quality. . But even if you use the G format to save a picture with a rich color of 100:::6.8Xl〇24 resolution, the compression rate and other secrets also need about 12GK of storage space, if you increase the pressure = or reduce the picture Resolution and segmentation of the image will result in a large loss of image courtesy and integrity. Ming Mai's reading of a picture is a flow of image compression technology of traditional skills: Figure. This process begins by expanding, first providing - the original image (1), , sitting by digital sampling and then performing a mathematical conversion S12, such as the town (five) coffee = ln! transf) conversion, and then, after the mathematical conversion of the data, After processing S13 'encoding the result of the quantization process', after encoding, the age ψ [51 /A i^r> bundles the entire process. Turning out the image I shrinks the data S15, the knot 1281642. In recent years, the image compression technology of wavelet transform is a field of current research. It can achieve better image waste than the traditional Na:::. Please refer to the second diagram, which is a flow chart of the wavelet transform image of the prior art. This process starts from the beginning, providing - Image 2/21 «I wavelet transforms the image data in a tower decomposition manner, and then 'converts the data to EZW (Embedded Zero-tree avelet Encoding) ^ SPIHT ( Set Partitioning In erarchical Trees Coding) method to quantify S23, then? The result of the word quantization is outputted after the image is compressed, and the image is compressed. S24 ' Finally, the entire flow is ended. Read the moon tea. Figure ' is a flow chart of the wavelet transform image decompression technique of the prior art. This process starts from the beginning, provides an image compression = according to S3, then 'decodes the image compression data, performs inverse quantization S33 on the decoded data, and then inversely converts the result of the inverse quantization into wavelet transform. S34, then, after the inverse conversion, the image data S35 is output, and the entire flow is ended. Please refer to the fourth figure, which is a schematic diagram of the wavelet transform of the prior art. Taking a 4x4 pixel original image 41 as an example, the wavelet transform performs a horizontal split 42 and a vertical split 43 on the original image 41. First, the horizontal segmentation 42 is performed. The first column of the original image 41 is labeled as A, β, c, and D. The first column of the horizontal segment 42 is A, B, C, and D. And Α, is α + ° β, β, is C + D, C is Α - Β and D is cD 'The same second column also has Ε, FH Η, subgraph ' ,, "^, sub-picture is divided into low frequency 1281642
L子圖區的部份，c，、d，、g，、h，HPart of the L subgraph area, c, d, g, h, H
=的部份。接著，進行垂直分割42，係、由水平 中以A，:E:獲得A，，、以A，_EW 以C G獲得D，，，而A”子圖區分為ll子圖區部份 低頻L子圖區中的低頻部份，β，，子 ：為 部份是為低頻L子圖區中的高頻部份，c”子圖J為^ 子圖區部份是為高頻Η子®區中的低頻部份，D，，子圖 分為HH子圖區部份是為高頻η子圖區中的高頻部^ 如此，區分出4x4像素的原始圖像41中LL、LH、HL及 冊各子圖區部份，即完成一階的小波轉換。換言之， LL、LH、HL及ΗΗ即是原始圖像41的子圖區。 立接續請參閱第五圖，係為習知技藝之塔式分解之示 意圖。圖中’包含一第一階小波轉換5卜一第二階： 波轉換52及-第三階小波轉換53。當進行第—階小波 ，換51後，產生LU、LH1、HL1及ΗΗ1的子圖區，而 第一階小波轉換52則是進一步的對LL1低頻的子圖區 進2小波轉換，產生LL2、LH2、HL2及HH2的子圖區， 接著，執行第三階小波轉換53即是對lL2低 區進行小波轉換，產生LL3、LH3、HL3及刪的= 如此，便完成三階小波轉換的塔式分解，並具有LL3、 LH3、HL3、HH3、LH2、HL2、HH2、LH1、HL1 及腿的 子圖區，其中，LL3是整個圖像最低頻的部份，由此一 子圖區可得整個圖像的輪廓。一般而言，將進一步區分 各子圖區能量占整個圖像資料能量的百分比以進行後 1281642 續的量化。 …由於上述基於小波轉換的新一代圖像壓縮技術可 以付到比傳統〗PEG標準更好的圖像處理效果，它已經 被納入了】PEG2_標準正在推廣應用。但由於可攜式 通=衣置處理益之卫作頻率較低，而傳統小波轉換是使 用塔式分解的方式，並且接著使用謂或spiHT的量化 方式，雖然可取得較好的圖像壓縮效果，但效能無法搭 配可攜式通訊裝置之系統限制。 有鐾於f知技藝之各項問題，實為廣大民眾所殷切 盼主及本發明人念茲在茲者，為了能夠兼顧解決之，本 毛明人基於多年從事圖像處理的研究開發與諸多實務 經驗，乃思及改良之意念，窮其個人之專業知識，並且 經多方研究設計與專題探討，遂於本發明係提出一種圖 像壓縮/解壓縮方法及其圖像編解碼系統以作為上述期 望一實現方式與依據。 【發明内容】 本發明之主要目的係揭露—種圖像壓縮方法。此方 法包含提供-如數位圖像之圖像，並透過—區域選擇單 兀’於圖像上選取一局部範圍，同時，更提供一第一編 碼程序，壓縮整個圖像以產生—第—圖像_數據，且 提供-第二編碼程序，I 縮所選取的局部範圍之圖像以 ^生21像_數據，再結合第—圖彻縮數據及 弟二圖㈣數據’以輸出一圖㈣縮資料。另外，更提 ⑧ 8 1281642 供：編碼演算程序作為上述之第一編碼程序及第二編 序’且第—編碼程序之—轉換階數係低於第二編碼 ^序之^數。以實現圖像較重要部份以較高的圖像壓 ^數據量壓縮’而不重要部份則以較低的圖像壓縮數據 里H達成於有限的傳輸量内傳遞圖像，以呈現較佳 的圖像效果。 本發明更揭露上述之編碼演算序，直 整個圖像之-平均w以下舰學式計^ Εϋ=^ΣΣ\ηυ)\2 其中，/(/，y·)係為圖像中每一個像素的能量、/= ι〜从 以及/ = 1〜Λ"。 八解敕，、小波轉換，係對應前述之轉換階數以分別 個圖像及選取之局部範圍之圖像，於結東該小波 Ί Μ別產生-最低頻子圖區及至少—子圖區， 二二瑕低頻子圖區係進行-類IPEG編碼。其於子圖 ^ 、圖£之一平均能量弋與整個圖 :之千：…進行比對’以判別子圖區是為一放棄狀 ^:待^碼狀態或-再執行一階小波轉換狀態其中之 :為-不重複之自然:數，上述之判斷式描述如下： 則判別該子圖區係為 轉換狀態； & 4 =，叫。’則判別該子圖區係為待編碼狀態； /叫，則判別該子圖區係為放棄狀態； 1281642 ^其中，且0<Cl<C2<1、、a係為-百分比。 ::i再執：r階小波轉換狀態之該子圖區後，係標示The part of =. Then, vertical segmentation 42 is performed, and A is obtained from A, E: E in the horizontal, D is obtained as CG in A, _EW, and D is divided into LV sub-pictures. The low frequency part of the picture area, β, and sub are: the part is the high frequency part of the low frequency L subgraph area, and the c" subgraph J is the part of the subgraph area for the high frequency dice® area. The low frequency part, D, and subgraph are divided into HH subgraph area, which is the high frequency part in the high frequency η subgraph area. Thus, the original image 41 of 4x4 pixels is distinguished from LL, LH, HL. Part of the sub-picture area of the book, that is, complete the first-order wavelet conversion. In other words, LL, LH, HL, and ΗΗ are the sub-picture areas of the original image 41. Please refer to the fifth figure for the connection, which is the intention of the tower decomposition of the conventional technique. The figure 'contains a first order wavelet transform 5b a second order: a wave transform 52 and a third order wavelet transform 53. When the first-order wavelet is performed, after 51, the sub-picture areas of LU, LH1, HL1, and ΗΗ1 are generated, and the first-order wavelet conversion 52 is further converted into 2 wavelet transforms to the LL1 low-frequency sub-picture area to generate LL2. Sub-picture areas of LH2, HL2, and HH2, and then performing the third-order wavelet transform 53 is to perform wavelet transform on the low region of lL2, and generate LL3, LH3, HL3, and deleted = so that the tower of the third-order wavelet transform is completed. Decomposed and has sub-picture areas of LL3, LH3, HL3, HH3, LH2, HL2, HH2, LH1, HL1 and legs, where LL3 is the lowest frequency part of the entire image, so that a sub-picture area can be obtained The outline of the image. In general, the energy of each subgraph area will be further divided into the percentage of the energy of the entire image data for subsequent quantitation of 1281642. ...Because the above-mentioned new generation image compression technology based on wavelet transform can pay better image processing effect than the traditional PEG standard, it has been included] PEG2_ standard is being promoted and applied. However, due to the low frequency of portable communication, the traditional wavelet transform is a method of tower decomposition, and then using the quantization method of YES or SPIHT, although better image compression effect can be achieved. However, the performance cannot be matched with the system limitation of the portable communication device. There are some problems in the art of knowing, and it is really for the general public to look forward to the Lord and the inventor. In order to be able to solve the problem, Ben Maoming is based on research and development of image processing for many years. Practical experience, thinking and improving ideas, poor personal expertise, and multi-disciplinary design and topical discussion, the present invention proposes an image compression/decompression method and its image encoding and decoding system as the above Expect an implementation and basis. SUMMARY OF THE INVENTION The main object of the present invention is to disclose an image compression method. The method includes providing an image such as a digital image, and selecting a partial range on the image through the region selection unit ,, and further providing a first encoding program to compress the entire image to generate a first image. Like _data, and provide - the second encoding program, the image of the local range selected by I is reduced to 21 image_data, and then combined with the first graph and the second graph (four) data to output a graph (4) Shrink the data. In addition, 8 8 1281642 is provided: the coding algorithm is used as the first coding program and the second sequence described above, and the conversion order of the first coding program is lower than the second coding sequence. In order to achieve a more important part of the image, the image compression is performed with a higher amount of image compression. The unimportant part of the image is compressed in a lower image, and the image is transmitted in a limited amount of transmission. Good image effects. The present invention further discloses the above-mentioned coding calculation sequence, which is the average of the following images - the average w is less than the ship's formula ^ Εϋ = ^ ΣΣ \ η υ) \ 2 where / (/, y ·) is each pixel in the image The energy, /= ι~ from and / = 1 ~ Λ ". Eight solutions, wavelet transform, corresponding to the above-mentioned conversion order to separate images and selected local range images, in the knot of the wavelet Μ discerning - the lowest frequency sub-picture area and at least - sub-picture area The second-order low-frequency sub-picture region is subjected to I-like IPEG coding. It is the average energy 弋 of the subgraph ^, the graph, and the whole graph: the thousand: ... for comparison 'to discriminate the subgraph area is a abandonment ^: wait for the code state or - then perform the first order wavelet transform state Among them: the natural number of - no repetition, the above judgment formula is described as follows: Then it is determined that the sub-picture area is the conversion state; & 4 =, called. Then, the sub-picture area is determined to be in a state to be coded; /, the sub-picture area is discriminated as abandonment state; 1281642^ where, and 0 <Cl<C2<1, a is a percentage. ::i re-execution: after the sub-picture area of the r-order wavelet conversion state, the system is marked
態之子:狀悲後，再提供―_圖塊，以分割待編碼狀 =⑥區產生至少一圖塊，並計算圖塊之一如方差之 t值。更提供—分佈區間，係用以比對前述之數值，以 “不圖塊是為一不重要圖塊、一一般重要圖塊、一較重 要圖塊或-重要圖塊其中之―’並依據該分佈區間之一 、為碼貝料1對圖塊進行—圖塊編碼，其中 間係為一數值分佈，如下所示： 刀佈L < ί] < t2 其中，^)、^及/2係為一旗標值； 田如方差之數值小於?〇，則標示該圖塊為不重要 圖塊； 、‘如方差之數值介於ί0及^間，則標示該圖塊為一 般重要圖塊； 當一如方差之數值介於（及間，則標示該圖塊為較 重要圖塊； 备一如方差之數值大於?2，則標示該圖塊為重要圖 塊。 最後，結合最低頻子圖區之類JPEG編碼及圖塊編碼， 產生第一圖像壓縮數據及第二圖像壓縮數據。 然而，第一編碼程序之一轉換階數係低於第二編碼 程序之一轉換階數，且轉換階數之需求係可自動設定咬 根據使用者自行設定。 1281642 :發明之另一目的係揭露一種圖像解壓縮方法。此 碼圖像壓縮資料’並區分出產生自一如編 鮮、H 弟一圖像壓縮數據及-第二圖像壓縮 严二二時’更提供一第一解碼程序，係針對第-圖像 據進行解壓縮，以產生一圖像，且提供一第二解 ^ 針對第二圖像壓縮數據進行解壓縮，以產生 :局：,圍之圖像。其後’結合圖像與局部範圍之圖 1 t山—平滑處理程序以消除兩者間之邊緣效應。 結合後之一如數位圖像之圖像。另外，更提 碼程：里二t碼旦'序作為上述之第-解碼程序及第二解 逆作用。“b里解碼程序係為前述之編碼演算程序之- 统。ΓίΓί之目的’本發明更揭露—種®像編解碼系 中，由=圖像編碼模組及—圖像解碼模組，其 ::且仃::以產生一圖像壓縮資料。另外，由圖像 針對圖像壓縮資料係進行解碼，並執行一 邊缘3 消除解碼後，圖像之局部範圍與圖像之 γ效應。此外’此系統更可以適用於一可攜式通訊裝 成之更對本發明之技術特徵及所達 實施例及相關圖式二下文謹提供較佳之 配合說明如後為辅佐之用，並以詳細之說明文字 1281642 【貫施方式】 請參閱第六圖，係顯示本發明之圖像壓縮 =圖。此方法之流程由開始展開，繼之，提供 : 圖像之圖像S61’並透過—區域選擇單元，於圖像上選 ^-局部範圍·，同時，提供—第—編碼程序 :個圖像以產生-第-圖像壓縮數據咖，更提供一第 二編碼程序，壓縮所選取的局部範圍之圖像以產^ =像壓^數據S64，再結合第—圖像壓縮數據及第二 圖_數據’以輸出_圖像壓縮資料咖。最後，結 此流程。其中/更提供—編碼演算程序作為上述之第一 編碼程序及第二編碼程序，而編碼演算程序將詳述於第 七圖。 第七圖，係顯示本發明所露揭之編碼 之“呈圖。此編碼演算程序由開始展開，首 $ ’計算圖像之-平均能量柳，再者，提供一小波 換’係對應-轉換階數以分別分解圖像及其局部範 圍’於結束削、波轉換後，以分職生—最低頻子圖區 =2子圖區’且最低頻子圖區係進行一類Jpeg編 =72。繼之，以—判斷式，將最低頻以外之子圖區之 平均能量圪與該平均能量五。進行比對，以判別子圖區 狀態、—待編碼狀態或—再執行—階小波轉換 狀，擇一，且⑺係為一不重複之自然數S73，於結束再 :了 P“、4波轉換狀態之子圖區後，係標示為待編碼狀 怨4。接著，提供一 ηχη圖塊，以分割待編碼狀態之 ⑧ 12 1281642 子圖區產生至少一圖塊，並計算圖塊之一數值S75，更 提供一分佈區間，係用以比對數值，以標示圖塊為一不 - 重要圖塊 般重要圖塊、一較重要圖塊或一重要圖 / 塊擇，並依據分佈區間之一編碼資料量對圖塊進行一 •圖塊：碼S76。其後’結合類JPEG編碼及圖塊編碼， 產生第圖像壓縮數據及第二圖像壓縮數據S77。最 後’結束整個流程。 明S閱弟八圖，係藏示弟六圖所述之判斷式產生之 示意圖。圖中，於獲得—圖像以f知塔式分解之三階小 波轉換81，接著，計算最低頻以外之子圖區之一平均 能量弋與整個圖像之平均能量尽進行比對，經由下列判 斷式： 若5m > Q X %，則判別子圖區係為一再執行一階小波 轉換狀態； 若弋>Clx尽，則判別子圖區係為一待編碼狀態； φ 若心< q X £。，則判別子圖區係為一放棄狀態； 其中，且0<W1 ’ c丨、C2係為一百分比，所為一不 重複之自然數係以標示各個子圖區。 於再執行-階小波轉換狀態之子圖區完成一次小波轉 ' ^後，即得一能量自適應分解之小波轉換82 ,此時， • 能量自適應分解之小波轉換82中之各子圖區僅存待編 碼狀態及放棄狀態，於圖中僅標示放棄狀態之子圖區。 緊接著’以一 ηχη圖塊分割，n可為一自然數，於此即 以一 4x4圖塊為例說明，計算各圖塊之—數值並以一分 13 1281642 佈區間比對數值標示圖塊，栌 知不為一不重要圖塊、一一 般重要圖塊二較重要圖塊或一重要圖塊，並依 應之一編碼資料量對圖塊進行一 曰、κ 仃圖塊編碼。而編碼資料 罝之分配於不重要圖塊即以2彳# ^ 、 圖塊則以2 + 16x1 + 1位亓；隹—站 力又置要 2+1fiv?4-9 ^ 疋進仃編碼、較重要圖塊則以 2+16X2+2位兀進偶碼、重要圖塊_ 2+16χ5 + 6 + 6位元進行編碼。進而眚银所^ 限的資料傳輸量内傳遞，僅之a 逆取於有 使之壬現較佳的圖像效果。 士請翏閱第九圖’係顯示本發明之圖像解壓縮方法之 4圖m之流程由開始展開，繼之，提供 壓縮資料S9卜並區分圖像壓縮資料之—第芦 數據及一第二圖像壓縮數據S92， ·象i縮 爆，冋時，提供一第一解 碼f序，解壓」縮第一圖像壓縮數據，以產生-圖像S93, 更^^供一弟·一解碼程序，解壓输黛-固你广 胂i細弟—圖像壓縮數據，以 ^生圖像之:局部範圍S94。接著，結合圖像與局部範 圍S95，再藉由一平滑處理藉床、由 处理私序以消除局部範圍與圖像 之邊緣效應S96 ’其後，輸出結合後之圖像挪。最後， 結束此流程。此處更提供—能量解碼轉作為上述之第 -解碼程序及第二解碼程序，此能量㈣㈣㈣第七 圖所述之編碼演算程序之一逆作用。 請-併參閱第十圖、第十—圖及第十二圖，係分別 顯不本發明之-種圖像編解碼系統、圖像編碼模级運作 及圖像解碼模組運作之示意圖。如第十圖巾所示，此圖 像編解碼系統1GG1包含—圖像編碼模組刚u及一圖 14 1281642 =碼模二刚:’其中’圖像編組運 圖所不，輸入-確立局部範圍之圖像1011，A十— 碼模組10011分別進行整個圖像編碼1()12 f像編 範圍編碼1〇13,其後結合兩者 俊，像局部 ㈣。另外’圖像解碼模組運作如第 1像壓縮資料1Q14,由圖像解碼模組_2 t 仃整個圖像解碼1021與圖像局部範 ： 處，序_以消除解瑪後，圖像』 I』像之邊緣效應，其後輸出—解碼圖像刪。此 卜此糸統更可適用於一可攜式通訊裝置。 離=所述僅為舉例性，㈣為限制性者。任何未脫 :明之精神與範疇，而對其進行之等效修改或變 更，均應包含於後附之申請專利範圍中。 【圖式簡單說明】 第：圖係為習知技藝之圖像壓縮技術之流程圖； =了圖係為習知技藝之小波轉換圖像壓縮技術之流程圖； ^圖係為*知技蟄之小波轉換圖像解壓縮技術之流程圖； 第四圖係為習知技藝之小波轉換之示意圖； ，五圖係為習知技藝之塔式分解之示意圖； 第六圖係顯示本發明之圖像壓縮方法之流程圖； ^七圖係顯示本發明所露揭之編碼演算程序之流程圖； 第八圖係顯示第六圖所述之判斷式產生之示意圖； 第九圖係顯示本發明之圖像解壓縮方法之流程圖； 1281642 第十圖係顯示本發明之一種圖像編解碼系統之示意圖； 第十一圖係顯示本發明之圖像編碼模組運作之示意圖；以及 第十二圖係顯示本發明之圖像解碼模組運作之示意圖。 【元件符號說明】 S11 -S15 ： 流程步驟； 41 ·· 原始圖像； S21 〜S24 ·· 流程步驟； 42 ： 水平分割； S31 〜S35 ·· 流程步驟； 43 ·· 垂直分割； A、] B、C及 D :像素； A，、 B， 、 C，、 D， 、 Ε， 、 F，、 G， 、 Η，、 A”、B”、C”及 L、] Η、LL、 LH 、 HL 及 ΗΗ :子圖區 • 51 ： 第一階 小波轉換； 52 : 第二階 小波轉換； 53 : 第三階 小波轉換；Son of the state: After the sorrow, provide the __block to divide the code to be coded =6 to generate at least one tile, and calculate one of the tiles such as the t value of the variance. Further, the distribution interval is used to compare the aforementioned values, so that "the non-block is an unimportant tile, a general important tile, a more important tile, or an important tile" One of the distribution intervals is code-blocking, and the block is coded, and the middle is a numerical distribution, as shown below: Knife L < ί] < t2 where ^), ^ and / The 2 series is a flag value; if the value of the field variance is less than 〇, the block is marked as an unimportant block; , 'If the value of the variance is between ί0 and ^, the block is marked as a general important map. Block; When the value of the variance is between (and between, it indicates that the tile is a more important tile; if the value of the variance is greater than ?2, the tile is marked as an important tile. Finally, the lowest frequency is combined. JPEG encoding and tile encoding such as sub-picture area, generating first image compressed data and second image compressed data. However, one of the first encoding programs has a lower conversion order than one of the second encoding programs. And the requirement of the conversion order can be automatically set according to the user's own setting. 1281642 Another object of the invention is to disclose an image decompression method. The code image compresses the data and distinguishes between the image generation data, the image compression data, and the second image compression. Further providing a first decoding program for decompressing the first image data to generate an image, and providing a second solution for decompressing the second image compressed data to generate: Bureau: The image of the surrounding. Then combine the image with the local range of Figure 1 t-smoothing to eliminate the edge effect between the two. The latter is combined with the image of the digital image. In addition, the code is more : The second two codes are used as the first-decoding program and the second inverse function. The decoding program in b is the aforementioned encoding algorithm. The purpose of the invention is to disclose the image of the image. In the codec system, the image coding module and the image decoding module are: and::: to generate an image compression data. In addition, the image is decoded for the image compression data system, and Perform an edge 3 to eliminate the local range of the image and the image γ after decoding In addition, the system can be applied to a portable communication device, and the technical features of the present invention and the related embodiments and the related drawings are provided below. Detailed explanation text 1281642 [Comprehensive method] Please refer to the sixth figure, which shows the image compression=graph of the present invention. The flow of this method is started from the beginning, and then, the image of the image S61' is transmitted through - The area selection unit selects a partial range on the image, and at the same time, provides a -first encoding program: an image to generate a -first image compression data, and a second encoding program to compress the selected The image of the local range is generated by the image compression data S64, and then combined with the image compression data and the second image data to output the image compression data. Finally, conclude this process. The code calculation program is provided as the first coding program and the second coding program, and the coding calculation program will be described in detail in the seventh picture. The seventh figure shows the "presentation of the code of the invention disclosed. The code calculation program starts from the beginning, the first $' calculates the image - the average energy willow, and further, provides a wavelet for the 'correspondence-conversion The order is to respectively decompose the image and its local range 'after the end of the cut, wave conversion, to divide the occupation - the lowest frequency sub-picture area = 2 sub-picture area ' and the lowest frequency sub-picture area to perform a class of Jpeg = 72. Then, by using the -judging formula, the average energy 圪 of the sub-picture area other than the lowest frequency is compared with the average energy five to determine the state of the sub-picture area, the state to be coded, or the re-execution-order wavelet transform. Alternatively, (7) is a non-repetitive natural number S73, and at the end of the P:, 4 wave conversion state of the sub-picture area, is marked as to be coded. Next, an ηχη tile is provided to divide at least one tile in the 8 12 1281642 sub-picture region to be encoded, and calculate a value S75 of the tile, and provide a distribution interval, which is used to compare the values to indicate The block is a non-critical block-like important block, a more important block or an important picture/block selection, and the block is coded according to one of the distribution intervals: a block: code S76. Thereafter, combined with JPEG encoding and tile encoding, first image compressed data and second image compressed data S77 are generated. Finally, the entire process ends. Ming S read the eight pictures of the younger brother, which is a schematic diagram of the judgment formula produced by the six brothers. In the figure, the third-order wavelet transform 81 is obtained by decomposing the image, and then the average energy 之一 of one of the sub-picture regions other than the lowest frequency is compared with the average energy of the entire image, and the following judgment is made. Formula: If 5m > QX %, the sub-picture area is determined to repeatedly perform the first-order wavelet transform state; if 弋 >Clx is completed, the sub-picture area is determined to be a coded state; φ if the heart < q X £. Then, the sub-picture area is discriminated as a discarding state; wherein, 0 < W1 ′ c丨, C2 is a percentage, which is a non-repeating natural number to indicate each sub-picture area. After performing a wavelet transform '^ in the sub-image region of the re-execution-order wavelet transform state, an energy-adaptive decomposition wavelet transform 82 is obtained. At this time, each sub-graph region in the wavelet transform 82 of the energy adaptive decomposition is only The coded state and the abandoned state are stored, and only the subgraph area of the abandoned state is indicated in the figure. Immediately following 'dividing with a ηχη tile, n can be a natural number. Here, a 4x4 tile is taken as an example. The value of each tile is calculated and the tile is compared with a value of 13 1281642. , knowing that it is not a non-important tile, a general important tile, a more important tile or an important tile, and encoding one block and one κ block according to one of the encoded data quantities. The coded data is allocated to the unimportant block by 2彳#^, and the block is 2 + 16x1 + 1 bit; 隹- station force is set to 2+1fiv? 4-9 ^ The more important tiles are encoded with 2+16X2+2 bits into the even code and important blocks _ 2+16χ5 + 6 + 6 bits. In addition, the amount of data transmitted by the silver is limited, and only a is reversed to have a better image effect. Please refer to the ninth figure' for the process of displaying the image decompression method of the present invention. The flow of the image is started from the beginning, and then the compressed data S9 is provided and the image compression data is distinguished. Two image compression data S92, such as i shrinks, when ,, provides a first decoding f sequence, decompresses the first image compressed data to generate - image S93, more ^ ^ for a decoding Program, decompression and transmission - solid you 胂 i fine brother - image compression data, to ^ image: local range S94. Next, the image and the local range S95 are combined, and then a smoothing process is used to borrow the bed, and the private sequence is processed to eliminate the edge effect of the local range and the image S96', and then the combined image is output. Finally, end this process. Further provided herein - energy decoding is performed as the first-decoding program and the second decoding program described above, and this energy (4) (4) (4) is one of the inverse of the coding algorithm described in the seventh figure. Please refer to the tenth, tenth, and twelfth drawings, which are schematic diagrams showing the operation of the image coding and decoding system, the image coding mode operation, and the image decoding module of the present invention, respectively. As shown in the tenth drawing, the image encoding and decoding system 1GG1 includes an image encoding module just u and a picture 14 1281642 = code mode two just: 'where the 'image grouping map does not, input-establish local The image 1011 of the range, the A-code module 10011 performs the entire image encoding 1 () 12 f image encoding range encoding 1 〇 13, respectively, and then combines the two, like the local (four). In addition, the 'image decoding module operates as the first image compression data 1Q14, and the image decoding module _2 t 仃 the entire image decoding 1021 and the image partial fan: where, in order to eliminate the image, I" like the edge effect, then output - decode the image deleted. Therefore, the system can be applied to a portable communication device. The above is only an example, and (4) is a restrictive one. Any changes or modifications to the spirit and scope of the invention shall be included in the scope of the appended patent application. [Simple description of the diagram] The first diagram is a flow chart of the image compression technology of the conventional technique; = The diagram is a flow chart of the wavelet transform image compression technology of the conventional technique; ^The diagram is *Knowledge technology The flow chart of the wavelet transform image decompression technology; the fourth figure is a schematic diagram of the wavelet transform of the prior art; the fifth figure is a schematic diagram of the tower decomposition of the conventional art; the sixth figure shows the diagram of the present invention A flowchart of a compression method; a diagram showing a code calculation procedure disclosed in the present invention; an eighth diagram showing a schematic diagram of the generation of the judgment pattern described in the sixth diagram; Flow chart of image decompression method; 1281642 FIG. 11 is a schematic diagram showing an image coding and decoding system of the present invention; FIG. 11 is a schematic diagram showing the operation of the image coding module of the present invention; A schematic diagram showing the operation of the image decoding module of the present invention. [Description of component symbols] S11 - S15 : Process steps; 41 ·· Original image; S21 to S24 ·· Process steps; 42: Horizontal division; S31 to S35 ·· Process steps; 43 ·· Vertical division; A,] B , C and D: Pixels; A, B, C, D, Ε, F, G, Η, A", B", C" and L,] Η, LL, LH, HL And ΗΗ : subgraph area • 51 : first order wavelet transform; 52 : second order wavelet transform; 53 : third order wavelet transform;
LL卜 Lm、HU、HH1、LL2、LH2、HL2、HH2、LL3、LH3、 HL3及HH3 :子圖區； S61〜S65 :流程步驟； S71〜S77 :流程步驟； 81 :塔式分解之三階小波轉換； 8 2 ·能|自適應分解之小波轉換； S91〜S96 :流程步驟； 1012、1013 :編碼； 1001 :圖像編解碼系統；1014 :圖像壓縮資料； 10011 ··圖像編碼模組；1021、1022 :解碼； 10012 ··圖像解碼模組；1023 :平滑處理程序；以及 1011 :圖像； 1024 :解碼圖像。 16LL 卜 Lm, HU, HH1, LL2, LH2, HL2, HH2, LL3, LH3, HL3 and HH3: sub-picture area; S61~S65: flow step; S71~S77: flow step; 81: third order of tower decomposition Wavelet transform; 8 2 · Energy | Wavelet transform with adaptive decomposition; S91~S96: Flow step; 1012, 1013: Encoding; 1001: Image codec system; 1014: Image compression data; 10011 · Image coding mode Group; 1021, 1022: decoding; 10012 · image decoding module; 1023: smoothing program; and 1011: image; 1024: decoded image. 16