201118755 六、發明說明： 【發明所屬之技術領域】 本發明是有關於一種影像辨識登入系統及其方法，特 別是有關於一種具有動態影像辨識功能之系統及其方 法0 【先前技術】 一般電子裝置（如：電腦或手機）的作業系統，諸如： 電腦的微軟視窗作業系統、Linux、UNIX，或手機的微軟 Windows Mobile、Android…等，或電子裝置的應用軟體， 諸如：MSN、資料庫軟體、數位監視軟體、電玩軟體…等 在使用時，必需先經由認證程序驗證登入身份及密碼，才 能開始使用作業系統或應用軟體。 以微軟視窗作業系統而言，其在啟動後，若有必要輸 入認證資料，則由作業系統載入一認證程序，提供一認證 視窗晝面，視窗晝面上顯示一身份欄位及一密碼攔位。使 用者透過操作一輸入單元，例如：鍵盤、虛擬鍵盤或觸控 螢幕…等，用以將至少一 ASCII字元，例如A、B、C...1、 2、3…等逐一鍵入至身份攔位及密碼攔位，以組成身份及 密碼之認證資料，用以進行認證，待確認資料無誤，則進 一步執行一内部程式或其他介面，進而可開始使用此作業 系統及其應用軟體。 且目前最普遍使用之網路交易的安全管理方式是使用 201118755 2二舉凡魏卡及賴财上的^，約㈣密碼來做 f身=的辨識。然而，使用密碼的機制有其先天 迷、碼容易被偷窺及盜取，資料 ' ''/〇 事後追查相m i被料&子所複製， 法，：二來許多研究致力於發展新的身份辨識方 去以改善網路交易的安全。 經相當成孰，包括j物辨識系統的發展已 辨識機制皆6被開發出來。然而，這些生 站 之設備皆很昂貴且操作不易。在二由、，.·而 識是較為可行⑽法，㈣舰生物辨識系統中，臉部辨 端僅而要一部攝影機配合系統的軟體 '、乍，除了設備的成本遠低於指紋、語音 方法外，在操作上亦簡單許多。 树工膜等等辨識 作前Ϊ之臉部辨識，只於在單一靜止臉部晝面上 部匕心人士往往利用圖片或照片便能矇編過臉 _識糸統，對於臉部辨識系統實為—大挑戰，因此，發展 種能克服Jl賴點的臉部影細識系統，便成為—項非常重 要的問題。 【發明内容】 有鑑於上述先前技術之問題，本發明之一目的就是 在提供一種影像辨識系統及其方法，以解決單一查 入易被偽造的問題。 根據本發明之-目的’提出—種影像辨識系統，其 〜辨識系統包含-儲存模組、一動態擷取模組及一處 201118755 理模組，儲存模組預先存入複數個第一動態影像5動態 影像擷取模組係於使用者欲登入時擷取一第二動態影 像，處理模組係連接於儲存模組與動態影像擷取模組， 處理模組接收動態影像擷取模組取得之第二動態影像， 並讀取儲存模組之第一動態影像，比對第二動態影像與 第一動態影像後，得出一比對結果，處理模組依據比對 結果提供下一步動作。 其中，第一動態影像係為複數個連續靜止晝面組 成，可經由動態擷取模組預先擷取存入儲存模組。 其中，第二動態影像係為複數個連續靜止晝面組成。 其中，第一動態影像係為臉部動作、頭部動作及肢 體動作等其中之一種動作。 其中，處理模組中包含一比對模組，此比對模組對 第一動態影像及第二動態影像中的連續靜止畫面，依照 時間區段節取出靜止晝面，再將此靜止晝面作一特徵比 對，此特徵比對的方式可為一灰階比對法、特徵值比對 法及區域縮減比對法（Region Of Interest, ROI)等其他影 像比對法其中之一、或任兩個及兩個以上。 其中，儲存裝置係具有一關聯式資料庫，用以儲存 第一動態影像。 其中，動態擷取模組係為一攝影機，提供用以擷取 第二動態影像。 其中，處理模組更具有一顯示模組，此顯示模組輸 出一指示訊息，此指示訊息係要求使用者進行登入所需 201118755 ' 之動作，如指定頭部動作或臉部動作等，此顯示模組更 - 可將第一動態影像輸出供使用者參考。 根據本發明之再一目的，提出一種影像辨識登入方 法，係適用於驗證使用者之登入程序，先於影像擷取模 組預先建立複數個第一動態影像於儲存模組後，當使用 者欲進行登入動作時，影像擷取模組擷取使用者之第二 動態影像，並將此第二動態影像傳送至處理模組，處理 模組之比對模組以第一動態影像及第二動態影之連續靜 _ 止晝面進行比對得出比對結果，處理模組依據此比對結 果進行下一步動作。 其中，當比對結果符合時，下一步動作係為執行登 入動作，反之，當比對結果不符合，下一步動作則為選 擇性操作行為，如提供另一種登入方式等，或再一次執 行驗證程序，或執行關機動作。 承上所述，依本發明之影像辨識登入系統及其方法， I 其可具有一或多個下述優點： (1) 可解決只用單一臉部辨識晝面時，欲入侵者只擁 有原始用者之照片就可以登入的情況。 (2) 利用臉部或頭部的各個角度畫面，達成難以仿 冒的特徵，大大增強了登入時的安全性。 【實施方式】 以下將參照相關圖式，說明依本創作之薄式螺管線 201118755 圈之發電結構之實施例，為使便於理解，下述實施例中 之相同元件係以相同之符號標示來說明。 請參閱第1圖，其係為本發明之影像辨識登入系統之 架構圖。圖中，其影像辨識登入系統包含一動態影像擷取 模組10、一處理模組11及一儲存模組12，動態影像擷 取模組10係連接於處理模組11，處理模組11包含一比 對模組110，儲存模組12係連接於處理模組11，且處理 模組11進一步更可連接一顯示模組13。 請參閱第2圖，其係為本發明之影像辨識登入方法之 流程圖。此方法，係適用於驗證使用者之登入程序，包 括下列步驟： (S10)預先由動態影像擷取裝置10或其他來源建立 複數個第一動態影像，此第一動態影像係為臉部動作、 頭部動作及肢體動作其中之一，並將此些第一動態影像 存入儲存模組12中； (S20)於動態影像擷取模組10擷取一第二動態影 像，此第二動態影像係為臉部動作、頭部動作及肢體動 作其中之一，並將此第二動態影像傳送至處理模組11 ; (S30)處理模組11中包含比對模組110,此比對模組 110於預先存入儲存模組12之第一動態影像及來自於動 態影像擷取模組10之第二動態影像中依據時間區段提 取出複數個連續靜止晝面，並依據此些連續靜止晝面作 特徵比對，此特徵比對係為灰階比對法、特徵值比對法 及區域縮減比對法（ROI)等影像比對法其中之一、任兩個 201118755 ' 或兩個以上，經過此特徵比對後，得出比對結果； • (S40)處理模組11依據特徵比對後之比對結果，當 此比對結果為符合時，進行步驟（S50) ’當不符合時，進 行步驟（S60); (S50)當處理模組11確認比對結果符合時，則進行 電子系統登入程序或下一步的功能操作使用；以及 (S60)當處理模組11確認比對結果不符合時，可以 φ 進行步驟（20)，或是結束登入程序，或執行關機，或是 選擇性的提供其他驗證登入功能。 請參閱第3圖’其係本發明之第一實施例操作影像 辨識登入系統的示意圖。圖中，動態影像榻取模組1〇係為 設置於筆記型電腦上方之攝影機，處理模組1丨係為設置於 筆記型電腦内之中央處理單元(CPU)及其他協助運算元件， 儲存模組12係為設置於筆記型電腦内之硬碟，且以預先存 入複數個第一動態影像，處理模組11中包含比對模組11〇， 鲁此比對模組係為預先撰寫好之影像比對程式，於執行比對時 被載入至處理模組11執行，處理模組於此實施例中連接至 顯示模組13，此顯示模組13係為筆記型電腦上之液晶顯示 器。 /、 當使用者進行登入程序時，顯示模組13顯示指示訊 130 ’此指示息130告知使用者此為利用轉頭動作作為 識，並告知轉頭方向與轉頭時間，於此實施例中，轉頭方 為相對使用者之右邊方向，並必須於12秒内完成此動作 請參閱第4圖’其係本發明之第一實施例影像辨 201118755 圖’並請啊參_3圖，圖中，當動態影像擷 像i、，’ S /取使用者之第二動態影像後，將此第二動態影 模組11之比賴组iig進行比對，此比對模 像心#Λ取預先存人儲雜組12之複軸第—動態影 象裝置10擷取並傳送至處理模組η之第二 ===::π”施例中’此特徵比對利用4 能η 動態影像之第4秒晝面·第-動 芯二::：41及第一動態影像之第12秒畫面42, 2—Λ 第4秒晝面43、第二動態影像之第8秒晝 =4及第二動態影像之第12秒晝面45各別進行特徵比對: :實加例中，此特徵比對係為利用影像特徵值比對法 ^像特徵值比·彻特定演算法計算出-或複數個特徵 值’並分別作對應比對。 參閱第3圖及第4圖，第一動態影像之第4秒畫面之特 徵，彻、第一動態影像之第8秒晝面之特徵值410及第一 動態影像之第12秒晝面之特徵值42G，於圖中組成第一特徵 值集合46 ’且第二動態影像之第4秒畫面之特徵值430、第 -動態影像之第8秒畫面之特徵值梢及第二動態影像之第 12秒，面之特徵值45〇，於圖中組成第二特徵值集合47，於 ^實加例中’第—特徵值集合46與第二特徵值集合47比對 、·Ό果為符合，並送出至處理模組11，則處理模組11進行八 入動作。 且 凊參閱第5圖，其係本發明之第二實施例操作影像 辨識登入系統的示意圖。圖中，於前述之圖相同之部分不 再贅述，僅於相異之特徵敘述說明，當使用者進行登入程 201118755 序時，顯示模組13顯示指示訊息130,此指示訊息13〇告知 使用者此為利用轉頭動作作為辨識，並告知轉頭方向與轉頭 時間，於此實施例中，轉頭方向為相對使用者之左邊方向， 並必須於15秒内完成此動作，但使用者利用平面照片14欲 偽造使用者正面之攝影。201118755 VI. Description of the Invention: [Technical Field] The present invention relates to an image recognition login system and a method thereof, and more particularly to a system with dynamic image recognition function and method thereof. [Prior Art] General Electronic Device (eg computer or mobile phone) operating system, such as: Microsoft Windows operating system for computers, Linux, UNIX, or Microsoft Windows Mobile, Android, etc. for mobile phones, or application software for electronic devices, such as: MSN, database software, When using digital monitoring software, video game software, etc., you must first verify the login ID and password through the authentication program before you can start using the operating system or application software. In the case of Microsoft Windows operating system, if it is necessary to input authentication data after startup, an authentication program is loaded by the operating system to provide an authentication window, and an identity field and a password barrier are displayed on the window. Bit. The user inputs an input unit, such as a keyboard, a virtual keyboard or a touch screen, etc., to type at least one ASCII character, such as A, B, C...1, 2, 3, etc. into the identity one by one. The block and password block are used to form the authentication data of the identity and password for authentication. If the information is correct, the internal program or other interface is further executed, and the operating system and its application software can be started. At present, the most commonly used security management method for online transactions is to use the 201118755 2 two-way Wei and Kay's ^, about (four) password to do the identification of the body =. However, the mechanism of using passwords has its innate fascination, the code is easy to be sneaked and stolen, and the information '''/after the incident is traced to the mi material & sub-copy, law, two: many studies are devoted to the development of new identities Identify the parties to improve the security of online transactions. After a considerable success, the development of the identification system including the identification system has been developed. However, the equipment of these stations is expensive and difficult to operate. In the second, , and .. is more feasible (10) method, (4) in the ship biometric identification system, the face is only required to have a camera with the software's software, and the cost of the device is much lower than the fingerprint and voice. In addition to the method, it is much simpler in operation. The tree film and the like are recognized as the face recognition of the front face, and only in the face of a single static face, the person who is at the top of the face often uses the picture or the photo to edit the face. The face recognition system is actually - Great challenges, therefore, the development of a species that can overcome Jl's face-to-face system is a very important issue. SUMMARY OF THE INVENTION In view of the above problems of the prior art, it is an object of the present invention to provide an image recognition system and method thereof to solve the problem that a single search is easily forged. According to the present invention, an image recognition system is provided. The identification system includes a storage module, a dynamic capture module and a 201118755 management module. The storage module is pre-stored in a plurality of first motion images. The dynamic image capture module is configured to capture a second dynamic image when the user wants to log in, the processing module is connected to the storage module and the dynamic image capturing module, and the processing module receives the dynamic image capturing module. The second dynamic image, and reading the first dynamic image of the storage module, comparing the second dynamic image with the first dynamic image, and obtaining a comparison result, the processing module provides the next action according to the comparison result. The first dynamic image is a plurality of continuous static kneading components, which can be pre-fetched into the storage module via the dynamic capture module. The second dynamic image is composed of a plurality of continuous stationary surfaces. The first motion image is one of a face motion, a head motion, and a limb motion. The processing module includes a comparison module, and the comparison module extracts the static side of the first moving image and the second moving image according to the time segment, and then the stationary surface For a feature comparison, the method of comparing the features may be one of a grayscale alignment method, an eigenvalue comparison method, and a region ratio comparison method (ROI), or Any two or more. The storage device has an associated database for storing the first motion image. The dynamic capture module is a camera that provides a second motion image. The processing module further has a display module, and the display module outputs an indication message, which is required for the user to perform the operation of the 201118755 ', such as specifying a head motion or a facial motion. Module is more - the first motion picture output can be used for user reference. According to still another object of the present invention, an image recognition login method is provided, which is suitable for verifying a user's login procedure, and before the image capture module pre-establishes a plurality of first motion images in the storage module, when the user desires During the login operation, the image capture module captures the second motion image of the user, and transmits the second motion image to the processing module, and the comparison module of the processing module uses the first motion image and the second dynamic image. The continuous static _ 昼 进行 进行 进行 进行 进行 进行 进行 比 比 比 比 比 比 比 。 。 。 。 。 。 。 。 。 。 。 Wherein, when the comparison result is met, the next action is to perform a login action, and conversely, when the comparison result does not match, the next action is a selective operation behavior, such as providing another login mode, or performing verification again. Program, or perform a shutdown action. As described above, the image recognition login system and method thereof according to the present invention, I can have one or more of the following advantages: (1) It can be solved that when only a single face recognition is used, the intruder only has the original The user's photo can be logged in. (2) Utilizing various angles of the face or head to achieve features that are difficult to counterfeit, and greatly enhancing the security at the time of login. [Embodiment] Hereinafter, embodiments of the power generation structure of the thin spiral pipeline 201118755 circle according to the present invention will be described with reference to the related drawings. For ease of understanding, the same components in the following embodiments are denoted by the same reference numerals. . Please refer to FIG. 1 , which is an architectural diagram of the image recognition login system of the present invention. In the figure, the image recognition login system includes a dynamic image capture module 10, a processing module 11 and a storage module 12. The dynamic image capture module 10 is connected to the processing module 11, and the processing module 11 includes The storage module 12 is connected to the processing module 11 , and the processing module 11 is further connected to a display module 13 . Please refer to FIG. 2, which is a flowchart of the image recognition login method of the present invention. The method is applicable to verifying the login process of the user, and includes the following steps: (S10) a plurality of first motion images are previously created by the motion image capturing device 10 or other sources, and the first motion image is a facial motion, One of the head motion and the limb motion, and the first motion image is stored in the storage module 12; (S20) the second motion image is captured by the motion image capturing module 10, and the second motion image is captured. One of the facial motion, the head motion, and the limb motion, and the second motion image is transmitted to the processing module 11; (S30) the processing module 11 includes the comparison module 110, the comparison module The plurality of consecutive still faces are extracted according to the time segment in the first motion image pre-stored in the storage module 12 and the second motion image from the motion image capturing module 10, and according to the continuous static frames The feature comparison is one of the image comparison methods such as gray scale comparison method, eigenvalue comparison method and region reduction comparison method (ROI), and any two of 201118755 ' or more than two After this feature comparison, (S40) The processing module 11 compares the results according to the feature comparison. When the comparison result is consistent, the step (S50) is performed. 'When not, the step (S60) is performed; (S50) When the processing module 11 confirms that the comparison result is met, the electronic system login procedure or the next functional operation is performed; and (S60) when the processing module 11 confirms that the comparison result is not met, the step (20) may be performed by φ. , or end the login process, or perform a shutdown, or optionally provide other verification login functions. Please refer to FIG. 3, which is a schematic diagram of an operation image recognition login system according to a first embodiment of the present invention. In the figure, the dynamic image couching module 1 is a camera disposed above the notebook computer, and the processing module 1 is a central processing unit (CPU) and other auxiliary computing components disposed in the notebook computer, and the storage module is stored. The group 12 is a hard disk set in the notebook computer, and a plurality of first motion pictures are pre-stored, and the processing module 11 includes a comparison module 11〇, and the comparison module is pre-written. The image matching program is loaded into the processing module 11 when performing the comparison. The processing module is connected to the display module 13 in this embodiment. The display module 13 is a liquid crystal display on the notebook computer. . / When the user performs the login process, the display module 13 displays the indication message 130 'This indication information 130 informs the user that the use of the turning action as the knowledge, and informs the turning direction and the turning time, in this embodiment The turning head is in the right direction of the user, and must complete the action in 12 seconds. Please refer to FIG. 4, which is the image of the first embodiment of the present invention, and the image is 201118755. In the case of the dynamic image image i,, 'S / after taking the second motion image of the user, the comparison of the comparison group iig of the second dynamic shadow module 11 is performed, and the comparison image is captured. The multi-axis first-motion image device 10 of the pre-existing storage group 12 is captured and transmitted to the second ===::π of the processing module η. In this example, the feature is compared with the 4 energy η dynamic image. The 4th second, the first-moving core 2:::41 and the 12th second screen of the first motion picture 42, 2—the 4th second face 43, the 8th second of the second motion picture 昼=4 and The 12th second of the second motion image is compared with each other by the feature 45: In the actual addition example, the feature comparison is to use the image feature value comparison method. The value is calculated by a specific algorithm to calculate - or a plurality of eigenvalues ' and is correspondingly compared. Referring to Figures 3 and 4, the characteristics of the 4th second picture of the first motion picture, the first motion picture The feature value 410 of the 8th second face and the feature value 42G of the 12th second face of the first motion image form the first feature value set 46' in the figure and the feature value of the 4th second picture of the second motion picture 430, the feature value of the 8th second picture of the first motion picture and the 12th second of the second motion picture, the feature value of the face is 45〇, and the second feature value set 47 is formed in the figure, in the actual case When the first feature value set 46 is compared with the second feature value set 47 and the result is matched and sent to the processing module 11, the processing module 11 performs an eight-in operation. Referring to FIG. 5, the system is The second embodiment of the invention is a schematic diagram of the operation of the image recognition login system. In the figure, the same parts in the foregoing figures are not described again, and only the different feature descriptions are described. When the user performs the login process 201118755, the display module is displayed. 13 displays an indication message 130, which indicates that the message 13 is used to inform This is to use the turning action as the identification, and to inform the turning direction and the turning time. In this embodiment, the turning direction is opposite to the left direction of the user, and the action must be completed within 15 seconds, but the user utilizes The flat photo 14 is intended to falsify the frontal photography of the user.
請參f第6圖，其係本發明之第二實施例影像辨識 比對的示意圖，並請同時參閱第5圖，圖中，於前述之圖 相同之部分不再贅述，僅於相異之特徵敘述說明，於此 實施例中itb特徵比對利用5秒為a相區段擷取出第一動態 影像，第5秒晝面6〇、第一動態影像之第1〇秒晝面^及^ -動態影像之第15秒晝面62，與第二動態影像之第5秒晝 面63去第三動態影像之第1〇秒晝面64及第二動態影像之第 15秒旦面65各別進行特徵比對，於此實施例中，此特徵比 像特徵值比對法，此影像特徵值比對法利用特 疋心4出—或複數個特徵值，並分別作對應比對。 '閱第5圖及第6圖’第一動態影像 徵值600、第一動綠爭德 弟办旦面之特 動離旦增之笛：像第畫之特徵值610及第- 晝面之特徵值620,於圖中組成第-特徵 值集。 第二動態影像之第5秒晝面之特徵值63〇、第 輸之獅叫== 於此實施财，第二成第二賴值集合67, . ^ _ 寺徵值集3 67係利用一張固定昭片產 -特徵面之各別晝面特徵值皆為相同，第 合，並送出至處理模H寺徵t集合67比對結果明顯為不符 理模組11，聽理模la U觸為登入失敗，Please refer to FIG. 6 , which is a schematic diagram of image recognition comparison according to the second embodiment of the present invention, and please refer to FIG. 5 at the same time, and the same parts in the above-mentioned figures are not described again, and only differs. The feature description shows that in this embodiment, the itb feature comparison uses 5 seconds for the a-phase segment to extract the first motion image, the fifth second is 6〇, the first motion image is the first second, and the ^ - the 15th second face 62 of the motion picture, and the 5th second face 63 of the second motion picture to the 1st second face 64 of the third motion picture and the 15th second face 65 of the second motion picture The feature comparison is performed. In this embodiment, the feature is compared with the image feature value comparison method, and the image feature value comparison method uses the special feature 4 or a plurality of feature values, and respectively performs corresponding comparison. 'Reading Figure 5 and Figure 6' The first dynamic image levy value 600, the first action green competition, the German dynasty, the special movement of the dynasty to increase the flute: like the feature value of the first picture 610 and the first - face The feature value 620 constitutes a first feature value set in the figure. The characteristic value of the 5th second face of the second motion picture is 63〇, the lion of the first loser == is implemented here, and the second is the second set of values of 67, . ^ _ Temple sign value set 3 67 system use one Zhang Dingzhao production - the characteristic surface of each feature surface are the same, the first combination, and sent to the processing mode H temple sign t set 67 comparison results are obviously unqualified module 11, listen to the mode la U touch For login failure,
Γ !s T 11 201118755 則進行選擇性驗财«重複_行驗證程序。 請參閱第7圖，复 辨識登入系統的示意圖H發明之第三實施例操作影像 再贅述，僅於相異c:於前述之圖相同之部分不 10係為連接於桌上型電丨/*述說明’動態影像操取模組 設置於桌上型電娜影機，處理模組11係為 元件，儲存模組12係為^^疋_)及其他協助運算 模組於此實關中連置=上型電_之硬碟，處理 桌上型電腦上之外接液晶顯示器。 、係為 當使用麵行登絲序時，顯㈣組u顯抑示訊息 30，似日不訊息13〇告知使用者此為利用轉頭動作作為辨 識’並告知轉頭方向與轉頭時間，於此實施例中，轉頭方向 為相對使用者之右邊方向，並必須於30㈣完成此動作。 請參閱第8 ® ’其係本發明之第三實施例影像辨識 比對的不意圖，並請同時參閱第7圖，圖巾，於前述之圖 相，之部分不再贅述’僅於相異之特徵敘述說明，於此 實施例中，為加強其辨識成功率及安全性，採用五張靜止晝 面作為辨識’此特徵比對利用6秒為時間區段擷取出第一二 癌影像，第6秒畫面8〇、第—動態影像之第12秒畫面81、 第-動態影像之第18秒晝面82、第一動態影像之第24秒畫 及第動態影像之第30秒晝面84,與第二動態影像之 f 6秒晝面85、第二動態影像之第12秒晝面86、第二動態 影像之f 18秒晝面87、第二動態影像之第24秒晝面沾及 第一動恕影像之第30秒晝面89各別進行特徵比對，於此實 201118755 =對特徵值比對法’此影像特 別作對應H 〜料异出—或複數個特徵值，並分 參閱第7圖及第皆 β 徵值_、第—動二圖，=細象之第6秒畫面之特 動〜、衫像之第12秒晝面之特徵值810、第一 動=像之第18秒晝面之特徵值82()、第—動態影像之第 私旦面之特徵值83〇及第一動態影像之第3〇秒晝面之特 徵值840’於圖中組成第一特徵值集合9〇，且第二動態影像 之第6秒晝面之特徵值85〇、第二動態影像之第以秒晝面之 特徵值860、第二動態影像之第18秒晝面之特徵值87〇、第 二動態影像之第24秒晝面88之特徵值880及第二動態影像 之第30秒畫面之特徵值890，於圖中組成第二特徵值集合 91，於此實施例中，第一特徵值集合9〇與第二特徵值集合 91比對結果為符合，並送出至處理模組u，則處理模組^ 進行登入動作。 以上所述僅為舉例性，而非為限制性者。任何未脫 離本發明之精神與範_，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。 13 201118755 【圖式簡單說明】 第1圖係為本發明之影像辨識登入系統之架構圖； 第2圖係為本發明之影像辨識登入方法之流程圖； 第3圖係為本發明之第一實施例操作影像辨識登入系 統之示意圖； 第4圖係為本發明之第一實施例操作影像辨識比對之 不意圖， 第5圖係為本發明之第二實施例操作影像辨識登入系 統之示意圖； 第6圖係為本發明之第二實施例操作影像辨識比對之 不意圖， 第7圖係為本發明之第三實施例操作影像辨識登入系 統之示意圖；以及 第8圖係為本發明之第三實施例操作影像辨識比對之 示意圖。 14 201118755 【主要元件符號說明】 10 :動態影像擷取模組； 11 :處理模組； 110 :比對模組； 12 :儲存模組； 13 :顯示模組； 130 :指示訊息； 14 :平面照片； 40 :第一動態影像之第4秒晝面； 400 :第一動態影像之第4秒晝面特徵值； 41 :第一動態影像之第8秒晝面； 410 :第一動態影像之第8秒晝面特徵值； 42 :第一動態影像之第12秒晝面； 420 :第一動態影像之第12秒畫面特徵值 43 :第二動態影像之第4秒晝面； 430:第二動態影像之第4秒畫面特徵值； 44 :第二動態影像之第8秒晝面； 440 :第二動態影像之第8秒晝面特徵值； 45 :第二動態影像之第12秒晝面； 450 :第二動態影像之第12秒晝面特徵值； 46 ·•第一特徵值集合； 15 201118755 47 :第二特徵值集合； 60 :第一動態影像之第5秒晝面； 600 :第一動態影像之第5秒晝面特徵值； 61 :第一動態影像之第10秒晝面； 610 :第一動態影像之第10秒晝面特徵值； 62 :第一動態影像之第15秒晝面； 620 :第一動態影像之第15秒晝面特徵值 63 :第二動態影像之第5秒晝面； 630 :第二動態影像之第5秒晝面特徵值； 64 :第二動態影像之第10秒晝面； 640 :第二動態影像之第10秒晝面特徵值； 65 :第二動態影像之第15秒晝面； 650 :第二動態影像之第15秒晝面特徵值； 66 :第一特徵值集合； 67 :第二特徵值集合； 80 :第一動態影像之第6秒晝面； 800 :第一動態影像之第6秒晝面特徵值； 81 :第一動態影像之第12秒晝面； 810 :第一動態影像之第12秒晝面特徵值； 82 :第一動態影像之第18秒晝面； 820 :第一動態影像之第18秒晝面特徵值 201118755 • 83 :第一動態影像之第24秒晝面； * 830 :第一動態影像之第24秒晝面特徵值； 84 :第一動態影像之第30秒畫面； 840 :第一動態影像之第30秒晝面特徵值； 85 :第二動態影像之第6秒晝面； 850 :第二動態影像之第6秒晝面特徵值； 86 :第二動態影像之第12秒畫面； ® 860 :第二動態影像之第12秒晝面特徵值； 87 :第二動態影像之第18秒晝面； 870 :第二動態影像之第18秒晝面特徵值； 88 :第二動態影像之第24秒晝面； 880 :第二動態影像之第24秒畫面特徵值； 89 :第二動態影像之第30秒晝面； $ 890 :第二動態影像之第30秒晝面特徵值； 90 :第一特徵值集合； 91 :第二特徵值集合；以及 S10〜S60:步驟。 17Γ !s T 11 201118755 Then carry out selective checkout «repetition_ line verification procedure. Please refer to FIG. 7 , a schematic diagram of the complex identification login system. The third embodiment of the invention is described in the operation image, and only the difference c: the same part of the above figure is not connected to the desktop power supply/* The description shows that 'the dynamic image manipulation module is set in the desktop type electric camera, the processing module 11 is the component, the storage module 12 is the ^^疋_) and other assisted computing modules are connected in this real mode. = The hard disk of the upper type is used to handle the external LCD display on the desktop computer. When the face line is used, the display (4) group displays the message 30, and the message does not indicate that the user is using the head movement as the identification and informs the direction of the head and the head. In this embodiment, the direction of the turning head is opposite to the right direction of the user, and the action must be completed at 30 (four). Please refer to Section 8 ® ' for the image recognition comparison of the third embodiment of the present invention, and please refer to FIG. 7 and the towel. In the foregoing drawings, the parts will not be described again. The feature description shows that in this embodiment, in order to enhance the recognition success rate and safety, five static images are used as the identification. This feature comparison uses 6 seconds as the time segment to extract the first two cancer images. The 6 second screen 8〇, the 12th second screen 81 of the first motion picture, the 18th second picture 82 of the first motion picture, the 24th second picture of the first motion picture, and the 30th second picture 84 of the motion picture, The f 6 second face 85 of the second motion image, the 12th second face 86 of the second motion picture, the f 18 second face 87 of the second motion picture, and the 24th second face of the second motion picture The first 30th of the image of the forgiveness image is compared with each other. In this case, 201118755 = the eigenvalue comparison method 'This image is specially corresponding to H ~ material - or a plurality of eigenvalues, and refer to Figure 7 and the second-order β-value _, the first-moving two-picture, = the special action of the 6th second picture of the fine image~, the shirt image The characteristic value 810 of the 12-second surface, the characteristic value 82 () of the first motion = the 18th second image of the image, the characteristic value 83 of the first private surface of the first motion image, and the third parameter of the first motion image The feature value 840' of the second face is composed of the first feature value set 9〇 in the figure, and the feature value of the 6th second face of the second motion image is 85〇, and the feature value of the second second image of the second motion image is the second feature image. 860, the characteristic value of the 18th second face of the second moving image is 87〇, the characteristic value 880 of the 24th second face of the second moving image, and the characteristic value 890 of the 30th second picture of the second moving image are shown in FIG. The second feature value set 91 is formed. In this embodiment, the first feature value set 9〇 and the second feature value set 91 are matched and sent to the processing module u, and the processing module ^ is logged in. action. The above is intended to be illustrative only and not limiting. Any and all equivalent modifications and changes may be made without departing from the spirit and scope of the invention. 13 201118755 [Simple description of the drawings] Figure 1 is a structural diagram of the image recognition login system of the present invention; Figure 2 is a flow chart of the image recognition login method of the present invention; Figure 3 is the first of the present invention FIG. 4 is a schematic diagram of operation image recognition comparison according to a first embodiment of the present invention, and FIG. 5 is a schematic diagram of an operation image recognition login system according to a second embodiment of the present invention; FIG. 6 is a schematic diagram of the operation of the image recognition comparison system according to the second embodiment of the present invention, and FIG. 7 is a schematic diagram of the operation image recognition login system according to the third embodiment of the present invention; and FIG. 8 is the present invention. The third embodiment operates a schematic diagram of image recognition alignment. 14 201118755 [Key component symbol description] 10: Dynamic image capture module; 11: Processing module; 110: Comparison module; 12: Storage module; 13: Display module; 130: Instruction message; Photo; 40: 4th second face of the first motion picture; 400: 4th second face feature value of the first motion picture; 41: 8th second picture of the first motion picture; 410: first motion picture 8th second surface feature value; 42: 12th second image of the first motion picture; 420: 12th second picture feature value of the first motion picture 43: 4th second picture of the second motion picture; 430: The 4th second picture feature value of the second motion picture; 44: the 8th second picture of the second motion picture; 440: the 8th second picture feature value of the second motion picture; 45: the 12th second of the second motion picture昼Face; 450: 12th second facet feature value of the second motion picture; 46 ·• first feature value set; 15 201118755 47: second feature value set; 60: 5th second face of the first motion picture; 600 : the 5th second facial feature value of the first motion image; 61: the 10th second surface of the first motion image; 610: the first motion image The 10th second face feature value; 62: the 15th second face of the first motion picture; 620: the 15th second face feature value of the first motion picture 63: the 5th second face of the second motion picture; 630: The 5th second facial feature value of the second motion image; 64: the 10th second plane of the second motion image; 640: the 10th second facial feature value of the second motion image; 65: the 15th of the second motion image Seconds; 650: 15th second facial feature value of the second motion picture; 66: first feature value set; 67: second feature value set; 80: 6th second face of the first motion picture; 800: The 6th second facial feature value of the first motion image; 81: the 12th second surface of the first motion image; 810: the 12th second facial feature value of the first motion image; 82: the 18th of the first motion image Second ;; 820: 18th second of the first motion image 特征 feature value 201118755 • 83: 24th second of the first motion picture; * 830: 24th second eigenvalue of the first motion picture; 84 : the 30th second picture of the first motion picture; 840: the 30th second facial feature value of the first motion picture; 85: the 6th second picture of the second motion picture; 850 : the 6th second eigenvalue of the second motion picture; 86: the 12th second picture of the second motion picture; 860: the 12th second eigenvalue of the second motion picture; 87: the second motion picture number 18 seconds face; 870: 18th second face feature value of the second motion picture; 88: 24th second face of the second motion picture; 880: 24th second picture feature value of the second motion picture; 89: The 30th second face of the second motion picture; $890: the 30th second facet feature value of the second motion picture; 90: the first feature value set; 91: the second feature value set; and S10~S60: step. 17