生物特徵識別技術與方法

《生物特徵識別技術與方法》從模式識別的角度進行生物特徵識別的方法學研究，主要探討套用最廣泛的人臉、指紋、掌紋等相關的檢測、識別、融合技術與方法。《生物特徵識別技術與方法》著力於介紹生物特徵識別領域目前正在發展的一些新理論和新技術，研究的方法均屬於目前模式識別領域十分活躍的方法，包括支持向量機、AdaBoost、流行學習理論、非線性降維、增量學習、鬆弛標註、摩爾特徵提取與識別、以及子空間學習等機器學習方法。

基本介紹

• 書名：生物特徵識別技術與方法
• 作者：胡德文 陳芳林
• 出版社：國防工業出版社
• 頁數：383頁
• 開本：16
• 品牌：國防工業出版社
• 外文名：Biometric Technique And Method
• 類型：科學與自然
• 出版日期：2013年8月1日
• 語種：簡體中文
• 定價：86.00
• ISBN：9787118089523
• 裝幀：精裝

內容簡介

《生物特徵識別技術與方法》由國防工業出版社出版，作者是胡德文/陳芳林。

圖書目錄

第1章緒論 1
1.1生物特徵識別簡述 1
1.1.1常用的生物特徵識別技術2
1.1.2各種常用生物特徵識別技術的比較9
1.2生物特徵識別技術發展概述 10
1.2.1生物特徵識別技術的優勢10
1.2.2生物特徵識別技術的市場發展概況11
1.2.3我國生物特徵識別技術發展概述12
參考文獻13
第2章人臉檢測與跟蹤 16
2.1概述 16
2.2AdaBoost學習算法 19
2.2.1AdaBoost方法19
2.2.2分類器訓練思想21
2.2.3構造弱分類器22
2.2.4構造強分類器27
2.2.5分類器訓練流程28
2.3多層級聯分類器 29
2.3.1多層級聯分類器的分類方法29
2.3.2使用級聯分類器進行訓練30
2.3.3實驗結果34
2.4基於AdaBoost的膚色檢測新方法 38
2.4.1膚色檢測方法38
2.4.2膚色分布的分析39
2.4.3實驗結果42
2.5人臉跟蹤 44
2.5.1智慧型像素聚類目標跟蹤算法44
2.5.2簡單背景的目標跟蹤47
2.5.3人臉跟蹤52
參考文獻56
第3章基於流形學習的人臉識別 60
3.1生長型局部線性嵌入算法 60
3.1.1生長模型分析60
3.1.2生長型局部線性嵌入算法66
3.1.3GLLE對Isomap算法改進的啟發82
3.2噪聲流形學習與分析 83
3.2.1問題的提出84
3.2.2鄰域平滑嵌入算法88
3.3流形學習算法的套用 98
3.3.1人臉序列資料庫的建立與評測99
3.3.2人臉序列中的流形結構103
3.3.3基於外觀流形的動態視頻人臉識別105
3.3.4基於流形重構的單圖像人臉識別107
參考文獻109
第4章多姿態人臉識別 115
4.1基於保持數據近鄰信息的增量學習方法 115
4.1.1增量Laplacian Eigenmaps(LE)算法115
4.1.2仿真實驗118
4.2引入遺忘機制的ART2改進算法 122
4.2.1自適應共振理論簡介123
4.2.2ART1神經網路123
4.2.3ART2神經網路125
4.2.4ART2網路存在的問題與改進127
4.3逆轉錄ART3算法 132
4.3.1ART3神經網路132
4.3.2ART3改進算法——ReART141
4.3.3仿真實驗145
4.3.4ReART在多姿態人臉識別中的套用150
4.4圖像平均重構技術與多姿態人臉識別 152
4.4.1自動人臉識別中的圖像平均技術153
4.4.2從原始圖像到平均臉：加權圖像平均技術155
4.4.3從平均臉到原始圖像：重構臉的生成159
4.4.4一個識別示例：視頻人臉識別160
4.4.5仿真實驗163
參考文獻167
第5章多特徵指紋識別 171
5.1概述 171
5.1.1指紋識別背景介紹171
5.1.2指紋識別綜述172
5.2現場重疊指紋的分離與特徵提取 179
5.2.1問題闡述179
5.2.2估計初始方向場181
5.2.3分離重疊方向場182
5.2.4分離重疊指紋及特徵提取189
5.2.5奇異點信息的套用190
5.2.6實驗192
5.3從細節點恢複方向場及其套用 197
5.3.1基於模型的方向場表示197
5.3.2從細節點恢複方向場199
5.3.3恢複方向場套用於指紋識別205
5.3.4實驗209
5.4指紋奇異點檢測 212
5.4.1問題闡述212
5.4.2指紋的拓撲分析214
5.4.3DORIC 特徵及其在去除虛假細節點上的套用216
5.4.4利用全局信息選擇奇異點的最優組合220
5.4.5實驗223
5.5多特徵融合與快速比對 227
5.5.1問題闡述227
5.5.2多特徵指紋識別的比較研究227
5.5.3基於分級結構的指紋多特徵辨認234
參考文獻240
第6章掌紋掌脈及其融合識別技術 248
6.1 概述 248
6.1.1掌紋識別技術的研究現狀248
6.1.2靜脈識別技術的基本原理250
6.1.3靜脈識別技術的研究現狀251
6.2掌紋特徵提取算法研究 254
6.2.1基於Moiré特徵的掌紋特徵提取算法255
6.2.2套用景象匹配的掌紋識別方法263
6.3掌紋識別系統的設計和實現 275
6.3.1基於數位相機的掌紋識別系統275
6.3.2基於視頻攝像頭的掌紋識別系統278
6.3.3基於ARM開發板的嵌入式掌紋識別系統281
6.3.4基於掃瞄器的嵌入式掌紋識別系統284
6.3.5實驗288
6.4掌脈採集系統及識別算法 289
6.4.1掌脈採集系統的設備選取289
6.4.2掌脈採集系統設計291
6.4.3掌脈識別算法研究293
6.5掌紋掌脈融合識別技術 300
6.5.1掌紋掌脈融合識別技術的原理301
6.5.2掌紋掌脈融合採集儀309
6.5.3掌紋掌脈融合技術的性能評價311
參考文獻315
第7章人臉與掌紋識別的子空間特徵提取方法 321
7.1直接局部保持投影算法及其在人臉與掌紋識別中的套用 321
7.1.1線性鑑別分析和直接線性鑑別分析323
7.1.2流形學習的概念與局部保持投影算法324
7.1.3直接局部保持投影算法及其計算326
7.1.4實驗和結果328
7.2二維局部保持投影算法及其在人臉與掌紋識別中的套用 331
7.2.1二維主成分分析的思想333
7.2.2二維局部保持投影算法335
7.2.3二維局部保持投影算法跟二維主成分分析的關係337
7.2.4對二維局部保持投影算法的進一步分析338
7.2.5實驗與結果339
7.3新的核局部保持投影算法及其在人臉與掌紋識別中的套用 343
7.3.1核方法的理論基礎344
7.3.2已有的核局部保持投影算法347
7.3.3新的核局部保持投影算法框架347
7.3.4核局部保持投影算法348
7.3.5實驗和結果350
7.4基於矩陣的圖像特徵提取方法的圖嵌入理論框架及其套用 354
7.4.1基於矩陣特徵提取算法的圖嵌入理論框架355
7.4.22DPCA、2DLDA 和 2DLPP算法的基於圖嵌入理論框架
解釋356
7.4.3非監督鑑別投影和邊界Fisher分析算法的矩陣形式推廣360
7.4.4二維鑑別嵌入分析(2DDEA)算法362
7.4.52DDEA與2DLDA的關係364
7.4.6基於2DDEA的圖像識別方法365
7.4.7實驗366
7.5一種基於PCA/ICA的人臉和掌紋特徵層融合策略 369
7.5.1使用主成分分析和獨立成分分析進行特徵提取370
7.5.2利用PCA、ICA 進行特徵層融合373
7.5.3實驗結果373
參考文獻376

Contents

Chapter 1Introduction1
1.1Introduction of Biometrics1
1.1.1Usually Used Biometric Technique2
1.1.2The Comparing of the Usually Used Biometric Technologies9
1.2Introduction of the Biometric Developing 10
1.2.1The Dominance of Biometrics10
1.2.2The Marketing of Biometrics11
1.2.3The Biometric Developing in China12
References13
Chapter 2Face Detection and Tracking16
2.1Introduction16
2.2AdaBoost Learning Algorithms19
2.2.1AdaBoost19
2.2.2The Classifier Training Idea21
2.2.3Constructing Weak Classifier22
2.2.4Constructing Strong Classifier27
2.2.5The Training Procedure of Classifier28
2.3Cascade Classifier29
2.3.1The Cascade Classifier Method29
2.3.2Use Cascade Classifier for Training30
2.3.3Experiment and Result34
2.4Skin Tone Detector Based on AdaBoost38
2.4.1Skin Tone Detecting Method38
2.4.2Skin Tone Distribution39
2.4.3Experiment and Result42
2.5Face Tracking44
2.5.1Object Tracking Based on Intelligent Pixel Clustering44
2.5.2Object Tracing in Simple Background47
2.5.3Face Tracking52
References56
Chapter 3Face Recognition Based on Manifold Learning60
3.1Growing Locally Linear Embedding60
3.1.1Growing Neural Gas Model60
3.1.2Growing Locally Linear Embedding66
3.1.3Improving Isomap Based on GLLE82
3.2Noisy Manifold Learning and Analysis83
3.2.1The Problem Analysis84
3.2.2Neighborhood Smoothing Embedding88
3.3The Application of Manifold Learning98
3.3.1Building the Face Database99
3.3.2The Manifold Structure of Face Sequence103
3.3.3Face Recognition in Video Based on the Appearance
Manifold Learning105
3.3.4Face Recognition in Single Image Based on Manifold
Reconstruction107
References109
Chapter 4Multi-pose Face Recognition115
4.1Incremental Laplacian Eigenmaps (LE) by Preserving Adjacent
Information115
4.1.1Laplacian Eigenmaps115
4.1.2Simulation Experiment118
4.2An Improved ART2 Algorithm by Introducing the Forgetting
Mechanism122
4.2.1The Introduction of ART123
4.2.2ART1 Neural Network123
4.2.3ART2 Neural Network125
4.2.4The Problem of ART2 and the Improvement127
4.3ReART Based on the Retrograde Mechanism of NO132
4.3.1ART3 Neural Network132
4.3.2The Improvement of ART3——ReART141
4.3.3Simulation Experiment145
4.3.4The Application of ReART in Multi-pose Face Recognition150
4.4The Image Averaging and Re-establishing Technique152
4.4.1The Image Averaging Technique in Automatic Face
Recognition153
4.4.2Weighted Image Averaging Technique155
4.4.3Generating Re-established Faces159
4.4.4Demo：Face Recognition in Video160
4.4.5Simulation Experiment163
References167
ⅩⅦChapter 5Multi-feature Fingerprint Recognition171
5.1Introduction171
5.1.1Introduction of Fingerprint Recognition171
5.1.2Introduction of Traditional Method172
5.2Separating Latent Overlapped Fingerprint and Feature Extraction179
5.2.1Problem Description179
5.2.2Original Orientation Field Estimation181
5.2.3Separating Overlapped Orientation Field182
5.2.4Separating Overlapped Fingerprint and Feature Extraction 189
5.2.5Utilizing the Singular Information190
5.2.6Experiment192
5.3Reconstructing Orientation Field form Minutiae and Its Application197
5.3.1Model based Description of Orientation Field197
5.3.2Reconstructing Orientation Field from Minutiae199
5.3.3Utilizing the Reconstructed Orientation Field for Fingerprint
Recogntion205
5.3.4Experiment209
5.4Singular Point Detection212
5.4.1Problem Description212
5.4.2Topologically Analysis of Fingerprint214
5.4.3DORIC Feature and Removing Spurious Singular Points216
5.4.4Selecting the Best Combination Base on Global Information220
5.4.5Experiment223
5.5Multi-feature Fusion and Fast Matching227
5.5.1Problem Description227
5.5.2A Comparative Study of Multi-feature Fingerprint Recognition227
5.5.3A Hierarchical Structure for Fingerprint Identification234
References240
Chapter 6Palmprint and Palmvein Recognition and the Fusion Method248
ⅩⅧ6.1Introduction248
6.1.1Review of Palmprint Recognition Technique248
6.1.2Fundamental Theory of Palmvein Recognition250
6.1.3Review of Palmvein Recognition251
6.2Palmprint Feature Extraction254
6.2.1Palmprint Feature Extraction based on Moiré Feature255
6.2.2Palmprint Recognition Methods by Scene Matching263
6.3Designing Palmprint Recognition System275
6.3.1Palmprint Recognition System Base on Digital Camera275
6.3.2Palmprint Recognition System Base on Video Camera278
6.3.3Palmprint Recognition System Base on ARM Development
Board281
6.3.4Palmprint Recognition System Base on Video Camera Scanner284
6.3.5Experiment288
6.4Palmvein Recognition System289
6.4.1Palmvein Acquisition Device289
6.4.2Designing Palmvein Acquisition System291
6.4.3Palmvein Recognition Method293
6.5The Fusion of Palmprint and Palmvein Recognition300
6.5.1The Theory of Palmprint and Palmvein Fusion Technique301
6.5.2The Palmprint and Palmvein Acquisition Device309
6.5.3Performance Analysis311
References315
Chapter 7Subspace Feature Extraction Methods for Face and Palmprint
Recognition321
7.1Direct Locality Preserving Projection (DLPP)321
7.1.1Linear Discriminant Analysis and Direct Linear Discriminant
Analysis323
7.1.2Manifold Learning and Locality Preserving Projection 324
7.1.3Direct Locality Preserving Projection326
7.1.4Experiment and Result328
7.2Two-dimensional Locality Preserving Projection (2DLPP)331
7.2.1Two-dimensional Principal Component Analysis (2DPCA)333
7.2.2Two-dimensional Locality Preserving Projection (2DLPP)335
7.2.3The Relationship between 2DLPP and 2DPCA337
7.2.4The Deeper Analysis of 2DLPP338
7.2.5Experiment and Result339
7.3Kernel Locality Preserving Projections (KLPP)343
7.3.1Kernel Theory344
7.3.2The Former Kernel Locality Preserving Projections347
7.3.3The Framework of the Proposed KLPP347
7.3.4The Proposed KLPP348
7.3.5Experiment and Result350
7.4Two-dimensional Discriminant Embedding Analysis (2DDEA),354
7.4.1Framework for Matrix-based Feature Extraction355
7.4.22DPCA, 2DLDA and 2DLPP : Graph Embedding Analysis356
7.4.3The Matrix Based Generalization of Unsupervised Discriminant
Projection and Fisher Analysis360
7.4.4Two-dimensional Discriminant Embedding Analysis362
7.4.5The Relationship between 2DDEA and 2DLDA364
7.4.6Image Recognition Based on 2DDEA365
7.4.7Experiment and Result366
ⅩⅨ7.5The Feature Fusion of Face and Palmprint Based on PCA/ICA369
7.5.1Feature Extraction Based on PCA and ICA370
7.5.2The Feature Fusion Based on PCA and ICA 373
7.5.3Experiment and Result373
References376