黃漢邦臺灣大學：機械工程學研究所李立明Lee, Li-MingLi-MingLee2007-11-282018-06-282007-11-282018-06-282005http://ntur.lib.ntu.edu.tw//handle/246246/61086在現今的生活中，指紋辨識系統對於我們日常生活而言已經是相當安全。但是倘若有朝一日生物科技有突破性的進展，或是某個歹徒將使用者的手指頭切下來，複製或取得一枚認證指紋將會是一件容易的事。有鑑於此，我們將發展一個防偽指紋辨識系統來確保系統所擷取到的指紋乃直接從認證者的手指中取得，而非從一枚複製或砍下的指頭中取得。 防偽指紋辨識系統是一個結合指紋與靜脈等生物特徵的辨識器。當使用者將其手指放置在感應區內，系統將啟動第一階段檢測機制，用以獲取使用者的主要生物特徵--指紋，來確認輸入的指紋屬於一位已認證的使用者。待系統確定該枚指紋屬於某位認證者後，第二階段的靜脈比對將會被啟動。在比對靜脈時，系統會將所擷取的使用者靜脈與指紋所確認之認證者靜脈相比對，以確保系統所擷取來的指紋是從該認證者手指所擷取來，而非是從擷取自一枚套在入侵者手上的複製指紋，或是一只被截下的死體指頭。如果所擷取的指紋在第一階段指紋辨識就不符合，系統將直接回絕該使用者而不再做相關的比較。In general, the fingerprint verification systems are good enough for daily secure life. But if the biometric technology of duplicating fingerprints has a significant improvement, or a bandit cuts off the authorizer’s finger, the only fingerprint verification procedure is not safe any more. Therefore, we have to develop an Anti-Forgery Fingerprint Verification System to ensure the fingerprint pattern directly extracting from one’s finger but not a duplicated fingerprint mask. An AFFVS (Anti-Forgery Fingerprint Verifications System) is a verification system combined with the biometric indicators of both fingerprint and finger-vein patterns. When a user places his finger on the sensor, the system will first check the primal feature, fingerprint pattern, to ensure that the input fingerprint pattern belongs to a authorized user, and then AFFVS will compare the input finger-vein pattern with the authorized user’s finger-vein pattern to guarantee the captured fingerprint is really acquired from the authorized person rather than the forgery fingerprint mask. But if the AFFVS decides the input fingerprint pattern as an invader, it will directly reject the user and will not do any vein-comparing procedure.摘要 i Abstract ii Table of Contents iii List of Tables v List of Figures vi Chapter 1 Introduction 1 1.1 Motivation 1 1.2 Related Works 4 1.3 Objectives and Contributions 7 1.4 System Foundational Operations 9 1.5 Thesis Organization 10 Chapter 2 Background Knowledge 12 2.1 Wavelet Theory 12 2.1.1 Multiresolution Analysis (MRA) 13 2.1.2 Wavelet Transform in One Dimension 14 2.1.3 Wavelet Transforms in Two Dimensions 20 2.2 Principle Component Analysis (PCA) 22 2.2.1 Maximizing transformed variance of PCA 23 2.3 FSVMs 29 2.3.1 SVMs (Support Vector Machines) 31 2.3.2 FSVMs (Fuzzy Support Vector Machines) 37 Chapter 3 Image Preprocessing 41 3.1 Progressive Fingerprint Enhancement 41 3.1.1 Histogram Equalization 45 3.1.2 Compute Orientations 49 3.1.3 Smooth Orientations 51 3.1.4 Fingerprint Segmentation 53 3.1.5 Gabor Filter 55 3.2 Finger-vein Extraction 58 3.2.1 Vein Extraction 59 3.2.2 Vein Processing 61 3.2.3 Luminance Strength 63 3.2.4 Extracted Vein 66 Chapter 4 Feature Extraction and Recognition 68 4.1 Calibration 68 4.2 Registration Point Location 70 4.2.1 Fingerprint Registration Point Locating 70 4.2.2 Finger-vein Registration Point Derivating 74 4.3 Wavelet Feature Extraction 75 4.3.1 Feature Block 76 4.3.2 Feature Region 78 4.4 PCA Feature Selection 81 4.5 Probability-FSVMs 82 4.6 Maximum Likelihood Criterion 85 Chapter 5 Experimental Results 88 5.1 System Configuration 88 5.1.1 Hardware Architecture 88 5.1.2 Graphic User Interface 89 5.2 System Communication 92 5.2.1 Communication Between Microprocessor and Fingerprint Sensor MBF200 93 5.2.2 Communication Between PC and Microprocessor 94 5.3 Fingerprint Classification Results 97 5.3.1 Biometric System Lab Database 97 5.3.2 NTU Database 99 5.3.3 Fracture Fingerprint Images 102 5.4 Experiments of AFFVS 104 5.4.1 NTU Database 104 5.4.2 Online Testing 106 Chapter 6 Conclusions 108 6.1 Conclusions 108 6.2 Future Works 109 References 110 Chapter 2 Background Knowledge 12 2.1 Wavelet Theory 12 2.1.1 Multiresolution Analysis (MRA) 13 2.1.2 Wavelet Transform in One Dimension 14 2.1.3 Wavelet Transforms in Two Dimensions 20 2.2 Principle Component Analysis (PCA) 22 2.2.1 Maximizing transformed variance of PCA 23 2.3 FSVMs 29 2.3.1 SVMs (Support Vector Machines) 31 2.3.2 FSVMs (Fuzzy Support Vector Machines) 37 Chapter 3 Image Preprocessing 41 3.1 Progressive Fingerprint Enhancement 41 3.1.1 Histogram Equalization 45 3.1.2 Compute Orientations 49 3.1.3 Smooth Orientations 51 3.1.4 Fingerprint Segmentation 53 3.1.5 Gabor Filter 55 3.2 Finger-vein Extraction 58 3.2.1 Vein Extraction 59 3.2.2 Vein Processing 61 3.2.3 Luminance Strength 63 3.2.4 Extracted Vein 66 Chapter 4 Feature Extraction and Recognition 68 4.1 Calibration 68 4.2 Registration Point Location 70 4.2.1 Fingerprint Registration Point Locating 70 4.2.2 Finger-vein Registration Point Derivating 74 4.3 Wavelet Feature Extraction 75 4.3.1 Feature Block 76 4.3.2 Feature Region 78 4.4 PCA Feature Selection 81 4.5 Probability-FSVMs 82 4.6 Maximum Likelihood Criterion 85 Chapter 5 Experimental Results 88 5.1 System Configuration 88 5.1.1 Hardware Architecture 88 5.1.2 Graphic User Interface 89 5.2 System Communication 92 5.2.1 Communication Between Microprocessor and Fingerprint Sensor MBF200 93 5.2.2 Communication Between PC and Microprocessor 94 5.3 Fingerprint Classification Results 97 5.3.1 Biometric System Lab Database 97 5.3.2 NTU Database 99 5.3.3 Fracture Fingerprint Images 102 5.4 Experiments of AFFVS 104 5.4.1 NTU Database 104 5.4.2 Online Testing 106 Chapter 6 Conclusions 108 6.1 Conclusions 108 6.2 Future Works 109 References 110en-US指紋辨識靜脈辨識fingerprintfinger-veinwaveletverification systemthesis