200923847 九、發明說明： 【發明所屬之技術領域】 本發明涉及—種電腦周邊器件，特別涉及-種用於電 腦之眼睛防護警示裝置及眼睛防護警示方法。 【先前技術】 一長時間使用電腦(長時間盯著電腦之顯示器)對眼睛有 告已經成為公知常識。為避免因此患上眼疾，宜在使用電 腦-段時間後，離開電腦以緩解視疲勞。然而，人們往往 忘情於工作或學習中，不能及時離開電腦，導致眼疾出現 或加重。 【發明内容】 有鐾於此’有必要提供—種在電腦使用者長時間使用 電腦後及時提醒電腦使用者離開電腦之眼睛防護 及眼睛防護警示方法。 。 -種眼睛防護警示技，其包括影像獲轉組、影像 處理模組、統計模組及警示模組。該影像獲取模組用於獲 取，腦使用者之連續多幢臉部圖像。影像處理模組用於處 理每賴臉部圖像以獲得表徵電腦使用者眼部高度(上下眼 皮之間之距離)之連續複數特徵值。統計模組用於統 續硬數特徵值t大於歉值之特難讀目 用者眼睛張開之時間。逖千丄 又”于电細使 月㈣（w s於在電腦使用 開之時間滿足預定條件時發出馨示。 月、 -種眼睛防護警示方法，其包括以下步驟： 獲取電服使用者之連續多悄臉部圖像； 7 200923847 處理每幀臉部圖 連續複數特徵值； 像以獲得表徵電腦使用者眼部高度之 以獲 不0 統計該連續複數特徵值_大於預定值之特徵值之數目 得電腦使用者眼睛張開之時間；及 在電腦使用者眼睛關之時間滿足預定條件時發出警 月防轉示裝置通賴大於預定值之特徵值(表 示電腦使用麵料於張開㈣）進賴相騎電腦使 用者眼睛闕之時間，並在電腦者眼睛_之時間滿 足預設條件(制超時餅)畴終示，鱗_ 者離開電腦。 【實施方式】 請參閱圖1，本發明之眼睛防護警示裝置1〇〇可用於 電腦200。電㉟勘可以係臺式電腦或筆記本電腦，本實 施方式之電腦200為筆記本電腦，其包括本體2()2及轉二 s史置於本體202邊緣之液晶顯示器2〇1。 請-併參_ 2，眼睛防護警示裝置勘包括影像獲 取模組10、影像處理模組2〇、統計模組3〇及警示模組奶二 影像獲取模組10驗獲取電腦制者之連射祕部圖 像。影像處理模組2G用域_巾貞臉部圖像以獲得表徵電 腦使用者眼部高度之連續複數特徵值。統計模組3〇用於統 計連續複數特徵值中大於預定值之特徵值之數目以獲得電 腦使用者眼睛張開之時間。警示模組4〇用於在電腦使用者 眼睛張開之時間滿足預定條件時發出警示。 200923847 本實施方式之影像獲取模組ίο採用攝像頭，其嵌設於 電腦200之液晶顯示器2〇1之邊緣(如左上角）。如此，攝像 頭與電腦使用者之臉部正對，可獲取完整、背景少之臉部 圖像，利於後續影像處理模組2〇對臉部圖像進行處理。 影像處理模組20包括區域劃分器21(regi〇n divisi〇n unit)，邊緣偵測器22(edge detection unit)及邊緣逼近器 23(edge approximation unit)。區域劃分器21用於對電腦使 用者之臉·像進行區_分以得到電腦制者卿所在 之眼部區域。邊緣偵測器22用於對眼部區域進行邊緣偵測 以銳化電腦使用者眼部之邊緣。邊緣逼近器23用於對銳化 之眼部區域進行邊緣逼近以得到特徵值。 優選地，影像處理模組20還包括均值濾波器24(mean filter)。均值濾波器24用於對眼部區域進行均值濾波以利 於眼部區域之邊緣偵測。 可以理解’影像處理模組20、統計模組30及警示模 0 了以採用各種編程語言，如visual basic，c，編 成之功能模組’並存儲於電腦200之本體202内之硬碟等 存儲設備。而影像獲取模組10通過線材與電腦200連接(如 與電腦200之主板連接），從而與影像處理模組20建立連 2。當通過電腦200之輸入設備(如鍵盤或滑鼠)或獨立啟動 °又備（圖未不)啟動眼睛防護警示裝置10Θ時，影像獲取模組 1〇開始獲取電腦使用者之臉部圖像，影像處理模組2〇、統 计拉組30及警示模組40同時載入電腦200之動態存儲設 備(如記憶體)並在其内運行。 警示模組40可通過在液晶顯 9 200923847 示器2〇1彈出警告框或觸發電腦200之發聲設備發出鳴叫 2式發出警示，提醒電腦使用者離開電腦200，緩解視 疲勞。 批f然’也可以制獨立於電腦之智慧設備裝載影像處 里杈、、且20、統計模組30及警示模組4〇。 »月多閱圖2 ’本發明之眼睛防護警示方 驟： S01 .獲取電腦使用者之連續多賴臉部圖像；具體地， 眼睛防護警示裝置⑽啟動後，影像獲取模組1G開始以 3〇幅/私之速率獲取電腦仙者之連續乡巾貞臉部圖像 ，並經 由線材傳輪至電腦200之動態存儲裝置； S02:處理每齡部圖像以獲得表徵電腦使用者眼部高 度之連續複數特徵值(詳細步驟請參下文）； S03 .統計連續複數特徵值中大於預定值之特徵值之數 目以獲得電腦使用者眼睛張開之時間；可以理解，若電腦 使用者眼睛處於張開狀態，眼部高度將大於預定值(如2毫 米），如此，可通過統計模組3〇統計複數特徵值中大於預 定值之特徵值之數目獲得電腦使用者眼睛張開之時間；及 S 〇4 ··在電腦使用者眼睛張開之時間滿足預定條件時發 出警示；預設條件可由電腦使用者設定，參考之預設條件 為：在一段時間(如1小時），大於預定值之特徵值之數目與 該段時間内臉部圖像之總幀數之比大於08(即在i小時 中，0.8小時電腦使用者眼睛張開；其餘0 2小時，電腦使 用者處於短暫閉眼或眨眼狀態）；更加具體地’若滿足預設 200923847 條件，則警示模組40通過在液晶顯示器2〇1彈出警告框或 觸發電腦200之發生設備發出鳴叫之方式發出警^〆 請參閱圖3，步驟S02(處理步驟）包括以下子步驟： 5021 :對臉部11像進行區域劃分以得到電腦使用者眼 部所在之眼部區域；具體地，區域劃分器21採用域值法對 臉部圖像進行區域劃分以得到電腦使用者眼部所在之眼部 區域， 5022 :對眼部區域進行均值濾波；具體地，均值濾波 器24對眼部區域進行均值濾波，以濾掉眼部區域内之圖像 雜訊’避免邊緣偵測器22在銳化電腦使用者眼部邊緣之同 時銳化圖像雜訊，而引起邊緣逼近器23誤判(誤判圖像雜 訊為電腦使用者眼部邊緣）； 5023 :對眼部區域進行邊緣偵測以銳化電腦使用者眼 部之邊緣；具體地，邊緣偵測器22採用s〇bel運算元對眼 部區域之亮度分量(brightness component)進行邊緣偵測，以 銳化電腦使用者眼部之邊緣；及 5024 .對銳化之眼部區域進行邊緣逼近以得到特徵 值，可以理解，經過均值濾波及邊緣偵測後，眼部區域之 叙化邊緣為眼珠邊緣，因此，通過邊緣逼近器23作縱向逼 近，便可得到特徵值。 可以理解，圖像處理模組2〇並不限於本實施方式，可 知用其他可代替模組，如，邊緣偵測器22採用其他運算元 (如Prewitt運算元或Marr-Hildret運算元）。 眼睛防護警示裝置1〇〇通過對大於預定值之特徵值進 11 200923847 n十以統片電腦使用者眼睛張開 ΐ眼睛張開之___時瓣示，^= 月旬使用者離開電腦。 电 出專本發财已符合發明專利要件，爰依法提 兴凡孰以上所財僅為本發明之紐實施方式， 砂飾藝之人士’於援依本案發明精神所作之等 f多飾或，祕1應包含於町之_請專利範圍内。 【圖式簡單說明】 圖1為本糾魏雜㈣ 腦之立體示tg。 elia不裝 =為本翻較佳實齡仏目㈣防縣林置之功能模 =為本發明較佳實施方式之眼睛防護警示方法之流 置應用於電 圖 =圖為核8报封施料Μ晴防護警 【主要組件符號說明】 眼睛防護警示裝置 100 影像獲取模組 程 示方法之一個子 影像處理模組 區域劃分器 邊緣偵測器 邊緣逼近器 10 20 21 22 23 均值濾波器 統計模組 警示模組 電腦 液晶顯不器 本體 24 30 40 200 201 202 12200923847 IX. INSTRUCTIONS: [Technical Field] The present invention relates to a computer peripheral device, and more particularly to an eye protection warning device and an eye protection warning method for a computer. [Prior Art] It has become common knowledge to use a computer for a long time (a display that has been staring at a computer for a long time) to the eye. In order to avoid eye diseases, it is advisable to leave the computer to relieve visual fatigue after using the computer for a period of time. However, people often forget to work or study, and can't leave the computer in time, causing eye diseases to appear or worsen. SUMMARY OF THE INVENTION It is necessary to provide an eye protection and eye protection warning method for promptly reminding a computer user to leave the computer after the computer user has used the computer for a long time. . - Eye protection warning technology, which includes image transfer group, image processing module, statistical module and warning module. The image acquisition module is configured to obtain a plurality of consecutive facial images of the brain user. The image processing module is configured to process each facial image to obtain successive complex eigenvalues that characterize the height of the computer user's eye (the distance between the upper and lower eyelids). The statistical module is used to continuously read the time when the hard eigenvalue t is greater than the apology.逖 丄 ” 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 w w w w w w w w w w w w a face image; 7 200923847 processing a continuous complex feature value of each frame of the face image; the image is obtained by characterizing the height of the eye of the computer user to obtain a number of feature values of the continuous complex feature value _ greater than a predetermined value The time when the computer user's eyes are open; and when the computer user's eyes are closed, the warning device is issued, and the alarm-proof device is used to pass the characteristic value greater than the predetermined value (indicating that the computer uses the fabric to open (4)). The time of the user's eyes is smashed, and the preset condition (system time-out cake) is satisfied at the time of the computer's eyes. The scale _ leaves the computer. [Embodiment] Please refer to FIG. 1, the eye protection warning of the present invention. The device 1 can be used for the computer 200. The computer 35 can be a desktop computer or a notebook computer, and the computer 200 of the embodiment is a notebook computer, which includes a body 2 () 2 and a second history. The liquid crystal display 2 〇 1 on the edge of the body 202. Please - and _ 2, the eye protection warning device includes image acquisition module 10, image processing module 2 统计, statistical module 3 〇 and warning module milk 2 image acquisition The module 10 obtains the image of the computer's continuous shooting secret part. The image processing module 2G uses the domain image to obtain the continuous complex feature value of the computer user's eye height. The number of the characteristic values greater than the predetermined value in the statistical continuous complex feature value is obtained to obtain the time when the computer user opens the eye. The warning module 4 is used to issue a warning when the computer user's eyes are open for a predetermined time. 200923847 The image acquisition module ίο adopts a camera embedded in the edge of the liquid crystal display 2〇1 of the computer 200 (such as the upper left corner). Thus, the camera faces the face of the computer user, and the face with a complete background is obtained. The image is processed by the subsequent image processing module 2 to process the facial image. The image processing module 20 includes a region divider 21 (regi〇n divisi〇n unit), and the edge detector 22 (ed) Ge detection unit) and edge approximation unit 23. The area divider 21 is used to divide the face and image of the computer user into the eye area where the computer maker is located. The edge detector 22 It is used for edge detection of the eye area to sharpen the edge of the eye of the computer user. The edge approximator 23 is used for edge approximating the sharpened eye area to obtain the feature value. Preferably, the image processing module 20 Also included is a mean filter 24. The averaging filter 24 is used for averaging the eye region to facilitate edge detection of the eye region. It can be understood that the image processing module 20, the statistical module 30, and the warning mode are understood. 0 is a storage device that uses a variety of programming languages, such as visual basic, c, compiled functional modules and stored in the body 202 of the computer 200. The image acquisition module 10 is connected to the computer 200 via a wire (for example, connected to the motherboard of the computer 200) to establish a connection with the image processing module 20. When the eye protection warning device 10 is activated by an input device (such as a keyboard or a mouse) of the computer 200 or independently activated, the image acquisition module 1 starts to acquire a facial image of the computer user. The image processing module 2, the statistical pull group 30, and the warning module 40 are simultaneously loaded into and operated by a dynamic storage device (such as a memory) of the computer 200. The warning module 40 can alert the computer user to leave the computer 200 by issuing a warning on the LCD 2 200923847 indicator 2 〇 1 pop-up warning box or triggering the audible device of the computer 200 to relieve visual fatigue. It is also possible to install a video device independent of the computer's smart device, and 20, the statistical module 30 and the warning module 4〇. »More than a month 2 'The eye protection warning of the present invention: S01. Obtain a continuous image of the face of the computer user; specifically, after the eye protection warning device (10) is activated, the image acquisition module 1G starts with 3 The width/private rate is obtained as a continuous image of the face of the computer fairy, and is transmitted to the dynamic storage device of the computer 200 via the wire; S02: processing the image of each age to obtain the height of the eye of the computer user The continuous complex eigenvalues (see the following for detailed steps); S03. Count the number of eigenvalues of the continuous complex eigenvalues greater than the predetermined value to obtain the time when the computer user opens the eyes; it can be understood that if the computer user's eyes are open The height of the eye will be greater than a predetermined value (eg, 2 mm). Thus, the time of opening the eyes of the computer user can be obtained by the statistical module 3 〇 counting the number of characteristic values of the complex eigenvalues greater than the predetermined value; and S 〇 4 ·· The warning is issued when the computer user's eyes are open at the time when the predetermined condition is met; the preset condition can be set by the computer user, and the preset condition of the reference is: for a period of time ( 1 hour), the ratio of the number of characteristic values greater than the predetermined value to the total number of frames of the facial image during the period of time is greater than 08 (ie, in i hours, 0.8 hours of computer user eyes open; the remaining 0 2 hours, The computer user is in a short-closed or blinking state; more specifically, if the preset 200923847 condition is met, the warning module 40 issues a warning by popping up the warning frame on the liquid crystal display 2〇1 or triggering the device of the computer 200 to emit a tweet. Please refer to FIG. 3, step S02 (processing step) includes the following sub-steps: 5021: area division of the face 11 image to obtain an eye area where the computer user's eye is located; specifically, the area divider 21 adopts a domain The value method divides the facial image into regions to obtain the eye region where the eyes of the computer user are located, 5022: performs mean filtering on the eye region; specifically, the averaging filter 24 performs mean filtering on the eye region to filter Image noise in the eye area is removed. Avoiding the edge detector 22 sharpening the image noise while sharpening the edge of the computer user's eye, causing the edge approximator 23 to misjudge (false positive figure) Like noise is the edge of the computer user's eye; 5023: Edge detection of the eye area to sharpen the edge of the computer user's eye; specifically, the edge detector 22 uses the s〇bel operator to the eye The brightness component of the region performs edge detection to sharpen the edge of the computer user's eye; and 5024. Edge approximation of the sharpened eye region to obtain the eigenvalue, it can be understood that the mean filtering and edge After detection, the edge of the eye region is the edge of the eyeball. Therefore, the feature value can be obtained by longitudinal approximation by the edge approximator 23. It can be understood that the image processing module 2 is not limited to the embodiment, and it is known that other alternative modules, such as the edge detector 22, use other arithmetic elements (such as Prewitt or Marr-Hildret). The eye protection warning device 1 〇〇 passes the characteristic value greater than the predetermined value into the 11 200923847 n tenth of the computer user's eyes open ΐ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The electricity issued by the special account has been in line with the requirements of the invention patent, and the above-mentioned financial assets are only the implementation method of the invention. The person of the sand decoration art has made a lot of decoration in the spirit of the invention. Secret 1 should be included in the scope of the _ _ patent. [Simple description of the figure] Figure 1 shows the stereoscopic tg of the brain. Elia is not installed = the best for the age of the eye (four) the function of the anti-county forest set = the eye protection warning method of the preferred embodiment of the present invention is applied to the electric map = the picture is the nuclear 8 seal application Μ晴防警警 [Main component symbol description] Eye protection warning device 100 Image acquisition module method of display Sub-image processing module Area divider edge detector edge approximator 10 20 21 22 23 Mean filter statistics module Warning module computer LCD display body 24 30 40 200 201 202 12