基於稀疏算法的人臉識別

《基於稀疏算法的人臉識別》是2014年12月國防工業出版社出版的圖書，作者是徐勇。

基本介紹

• 書名：基於稀疏算法的人臉識別
• 作者：徐勇
• ISBN：978-7-118-09758-0
• 頁數：227
• 定價：78.0
• 出版社：國防工業出版社
• 出版時間：2014年12月

圖書信息

書名基於稀疏算法的人臉識別

書號978-7-118-09758-0

作者徐勇

出版時間2014年12月

譯者

版次1版1次

開本16

裝幀精裝

出版基金國防科技圖書出版基金

頁數227

字數284

中圖分類TP39.4

叢書名

定價78.00

內容簡介

本書重點介紹稀疏算法及其改進方法在人臉識別中的套用，共分三部分。第一部分介紹降維方法等經典人臉描述與識別方法。第二部分介紹“局部”人臉描述與識別方法，重點介紹套用於人臉識別的原始稀疏方法原理、後來發展的稀疏方法以及基於稀疏描述思想的常規方法的改進，分析該類方法的本質特點。

第三部分介紹彩色人臉識別、視頻人臉識別和廣義人臉識別範疇的人臉偽裝判識技術，以及自主研發的人臉識別系統。 本書既可供自動化、計算機、電子工程等專業研究人員使用，又可供模式識別、機器學習、計算機視覺和圖像處理等開發人員參考。 　This book focuses on sparse representation and its improvement with the appl ication on face recognition. It contains three main parts. The first part introd uces the typical face representation and recognition methods involving dimension ality reduction. The second part mainly presents local representation for face r ecognition. Specifically, this part introduces typical face recognition methods based on sparse representation and its variation, as well as the improvement of classica l recognition methods that are based on the basic idea of sparsity. The third pa rt introduces two aspects of the face recognition system. The first aspect inclu des color face recognition, video face recognition, and the face disguise recogn ition technique. The second one is the face recognition system we developed. This book not only is suitable for researchers in the fields of automation,compu ter and electronic engineering,but also is very helpful for practitioners in the pattern recognition,machine learning,computer vision and image Processing communities.

圖書目錄

第1章引論1

1.1概述1

1.2人臉辨識與人臉認證評價指標2

1.3人臉識別方法4

1.3.1基於幾何特徵的人臉識別4

1.3.2基於表象的人臉識別6

1.3.3基於稀疏描述的人臉識別方法7

1.4人臉識別技術的套用分析10

1.5基於表情的人臉識別12

1.6年齡不變人臉識別14

1.73D人臉識別研究15

1.7.1基於空域的直接匹配方法15

1.7.2基於局部特徵的匹配16

1.7.3基於全局特徵的匹配17

1.8常用人臉庫介紹17

1.8.1FERET人臉資料庫17

1.8.2Yale人臉資料庫18

1.8.3Yale B人臉資料庫19

1.8.4ORL人臉資料庫19

1.8.5AR人臉資料庫19

1.8.6XM2VTS人臉資料庫21

1.8.7CMU PIE資料庫 21

1.8.8可見光與近紅外人臉資料庫22

第2章一維降維方法與人臉識別26

2.1特徵臉方法26

2.2基於Fisher準則的線性鑑別分析方法30

2.3Fisherface33

2.4一維核方法35

2.5局部保持投影方法37

第3章二維降維方法與人臉識別40

3.1二維主成分分析的實現及融合方案40

3.1.12DPCA40

3.1.22DPCA的兩種不同實現及其意義41

3.1.3實驗及分析42

3.2基於復矩陣的主成分分析與線性鑑別分析453.2.1基於復矩陣的主成分分析45

3.2.2基於復矩陣的主成分分析的討論45

3.2.3基於復矩陣的線性鑑別分析46

3.2.4基於復矩陣的線性鑑別分析的理論分析48

3.3二維局部保持投影分析50

3.3.1二維監督的局部保持投影50

3.3.2二維監督局部保持投影分析51

3.3.3二維判別監督局部保持算法51

3.3.4實驗52

第4章稀疏描述及其人臉識別套用53

4.1基於稀疏描述的人臉識別方法53

4.2快速稀疏方法 58

4.3字典學習60

4.4人臉對齊66

4.5核稀疏方法68

4.6本章小結69

第5章快速稀疏描述方法70

5.1基於全局表達方法的圖像測試樣本描述與識別70

5.1.1原空間中圖像測試樣本的全局表達方法70

5.1.2特徵空間中圖像測試樣本全局表達方法的初步設計71

5.1.3全局表達方法的可行性分析73

5.2快速稀疏描述方法76

5.2.1快速稀疏描述方法的設計77

5.2.2快速稀疏描述方法的可行性分析78

5.3快速稀疏描述方法的變形算法82

5.4本章小結86

第6章稀疏描述思想與改進的K近鄰分類88

6.1基於描述的最近鄰分類89

6.1.1基於描述的最近鄰分類方法89

6.1.2方法的分析90

6.1.3實驗結果92

6.1.4結論95

6.2加權最近鄰分類96

6.2.1方法介紹96

6.2.2加權最近鄰分類與NNCM的關係96

6.2.3實驗結果98

6.2.4結論101

6.3改進的近鄰特徵空間方法102

6.3.1K近鄰分類方法的幾個擴展102

6.3.2改進的近鄰特徵空間方法103

6.3.3改進的近鄰特徵空間方法分析104

6.3.4實驗結果109

6.3.5結論111

第7章稀疏描述思想與改進的降維方法112

7.1稀疏描述與常規變換方法的結合112

7.1.1改進的常規變換方法113

7.1.2關於ICTM的分析114

7.1.3實驗結果116

7.1.4結論118

7.2基於描述和降維的人臉識別118

7.2.1基於描述和降維的人臉識別方法119

7.2.2方法合理性分析120

7.2.3實驗結果121

7.2.4結論122

7.3討論122

7.3.1正確分類的前提條件122

7.3.2線性降維與樣本近鄰關係123

7.3.3描述誤差與分類精度124

第8章基於描述的方法與多生物

特徵識別套用126

8.1基於交叉得分的多生物特徵識別方法127

8.1.1具體方法127

8.1.2算法特點與原理129

8.1.3實驗結果130

8.1.4結論和討論133

8.2基於多Gabor特徵融合的人臉識別133

8.2.1Gabor變換及本節的方法134

8.2.2方法的分析136

8.2.3實驗137

8.2.4結論139

8.3復空間局部保持投影方法140

8.3.1CLPP140

8.3.2實驗142

8.3.3結論145

第9章彩色人臉識別的研究與發展147

9.1全局彩色人臉識別方法147

9.1.1基於各顏色通道的全局方法147

9.1.2其他全局方法148

9.2局部彩色人臉識別方法152

9.2.1局部二元模式運算元152

9.2.2基於Gabor變換的局部彩色人臉識別方法153

9.3其他彩色人臉識別方法153

第10章基於視頻的人臉識別技術綜述156

10.1引言156

10.1.1基於視頻的人臉識別技術的特點157

10.1.2基於視頻的人臉識別技術的幾個問題157

10.1.3基於視頻的人臉識別技術流程圖159

10.2各模組的主要技術與方法159

10.2.1基於視頻的人臉檢測159

10.2.2基於視頻的人臉跟蹤161

10.2.3基於視頻的人臉識別162

10.3基於視頻的人臉識別技術難點164

10.3.1光照的影響164

10.3.2姿態變化的影響164

10.3.3低解析度的影響165

10.4基於視頻的人臉識別技術在ATM中的套用165

10.4.1基於視頻的人臉識別技術在ATM中套用的技

術方案165

10.4.2技術方案流程圖166

10.4.3人臉識別的融合算法介紹167

10.4.4實驗驗證167

10.5實驗設計與結果統計169

第11章人臉偽裝判識及套用172

11.1概述與分析173

11.1.1人臉偽裝判識的實際需求173

11.1.2相關研究介紹175

11.1.3技術難點分析176

11.1.4偽裝定義、偽裝判識的全局方案177

11.2人臉偽裝判識方法179

11.2.1人像定位179

11.2.2基於Haar分類器的偽裝檢測180

11.2.3對誤報的分類186

11.2.4基於直線檢測的背景誤報剔除方法188

11.2.5基於鏡橋檢測的墨鏡誤報判別方法189

11.2.6基於對稱差分的墨鏡偽裝判別方法191

11.2.7基於Gabor濾波的墨鏡誤報判別192

11.2.8基於聚類分析的帽子誤報判別195

11.2.9基於膚色模型的口罩偽裝判識200

11.2.10其他誤報剔除方法201

11.3系統的實施與評估201

11.3.1系統結構201

11.3.2系統模組設計202

第12章人臉考勤與識別系統206

12.1紅外人臉識別與雙模態人臉考勤系統206

12.1.1遠紅外與近紅外人臉對比分析207

12.1.2系統技術方案的考慮與設計207

12.1.3系統軟硬體208

12.2人臉與指紋聯合識別系統212

12.2.1系統分析與設計212

12.2.2系統硬體214

12.2.3系統細節215

12.3基於人臉圖像認證的計算機登錄系統217

12.3.1引言217

12.3.2Windows 登錄系統概述217

12.3.3人臉登錄系統設計218

12.3.4人臉識別220

參考文獻222

Contents

Chapter 1Introduction1

1.1Introduction1

1.2Evaluation of face recognition and authentication2

1.3Face recognition methods4

1.3.1Face recognition based on geometric features4

1.3.2Appearance based face recognition 6

1.3.3Face recognition based on sparse representation 7

1.4 Application analysis of face recognition10

1.5 Expression based face recognition12

1.6 Age-invarant face recogniton14

1.73D face recogniton15

1.7.1Direct match in space domain15

1.7.2Local feature match16

1.7.3Global feature match17

1.8 Face database17

1.8.1 FERET face database17

1.8.2 Yale face database18

1.8.3 YaleB face database19

1.8.4 ORL face database19

1.8.5 AR face database19

1.8.6 XM2VTS face database21

1.8.7 CMU PIE face database21

1.8.8 Visible light and near-infrared face database22

Chapter 2 1D dimensionality reduction and face recognition26

2.1Eigenface26

2.2Linear discriminant analysis (LDA) using Fisher criterion 30

2.3Fisherface33

2.4One-dimensional kernel method 35

2.5LPP37

Chapter 3 2D dimensionality reduction and face recogntion40

3.12DPCA implementation and fusion40

3.1.12DPCA algorithm40

3.1.2Two implementations of 2DPCA41

3.1.3Experiments42

3.2PCA and LDA based on complex matrix45

3.2.1PCA based on complex matrix45

3.2.2Discussion on complex matrix based PCA45

3.2.3LDA based on complex matrix46

3.2.4Theoretical analysis on LDA based on complex matrix48

3.3 2DLPP50

3.3.12D supervised LPP50

3.3.2Analysis of 2D supervised LPP51

3.3.3Algorithm of 2D supervised LPP 51

3.3.4Experiments52

Chapter 4Sparse representation and face recognition53

4.1Face recognition based on sparse representation53

4.2Fast sparse methods58

4.3Code learning60

4.4Face align 66

4.5Kernel sparse representation68

4.6Summary69

Chapter 5Fast sparse representation70

5.1Test image recognition based on global representation70

5.1.1Global reprentation of test image in original space 70

5.1.2Global reprentation design of test image in feature space71

5.1.3Feasibility study on global representation73

5.2Fast sparse representation method76

5.2.1Design of fast representation method77

5.2.2Analysis of fast representation method78

5.3 Reformulation of fast representation method82

5.4Summary86

Chapter 6 Sparse representation and improved KNN88

6.1Representation based nearest neighbor classification

89

6.1.1Representation based nearest neighbor classification algorithm89

6.1.2Analysis90

6.1.3experiments92

6.1.4Summary95

6.2Weighted nearest neighbor classification (WNNC)96

6.2.1Introduction96

6.2.2Relationship between WNNC and NNCM96

6.2.3Experiments98

6.2.4Conclusions101

6.3Improved nearest neighbor feature space102

6.3.1Extensions of KNN102

6.3.2Improved nearest neighbor featrue space algorithm103

6.3.3Analysis of improved nearest neighbor feature space104

6.3.4Experiments109

6.3.5Conclusions111

Chapter 7 Sparse representation and improved dimensionality reduction112

7.1Combination of sparse representation and transformation method112

7.1.1Improved conventional transformation method (ICTM)113

7.1.2Analysis of ICTM114

7.1.3Experiments116

7.1.4Conclusions118

7.2Representatioin and dimensionality reduction based face recognition118

7.2.1Algorithm119

7.2.2Analysis120

7.2.3Experiments121

7.2.4Conclusions122

7.3Discussion122

7.3.1“Premise” of correct classification122

7.3.2Relationship between linear dimensionality reduction and nearest neighor123

7.3.3Representation error and classification accuracy124

Chapter 8Representation based method and multiple biometric recognition 126

8.1Multiple biometric recognition using cross-score fusion127

8.1.1Algorithm127

8.1.2Characteristic and principle of algorithm129

8.1.3Experiments130

8.1.4Conclusion and discussion133

8.2Face recognition based on multi-Gabor feature fusion133

8.2.1Gabor transformantion and its two schemes134

8.2.2Analysis of two schemes136

8.2.3Experiments137

8.2.4Conclusions139

8.3Complex space LPP (CLPP)140

8.3.1CLPP algorithm140

8.3.2Experiments142

8.3.3Conclusions145

Chapter 9 Color face recognition147

9.1Global color face recognition147

9.1.1Global method based on all color channels147

9.1.2Other global methods148

9.2Local color face recognition152

9.2.1Local binary pattern (LBP)152

9.2.2Gabor transformation based local color face recognition153

9.3Other color face recognition methods153

Chapter 10 Review of face recogniton in video156

10.1Introduction156

10.1.1Characteristics of face recognition in video157

10.1.2Problems of face recognition in video157

10.1.3Flow chart of face recognition in video159

10.2Main techniques of modules159

10.2.1Face detection in video159

10.2.2Face track in video161

10.2.3Face recognition in video162

10.3Difficulties of face recognition in vedio164

10.3.1Influence of light164

10.3.2Pose variation164

10.3.3Influence of low resolution165

10.4Applications of vedio based face recognition in ATM165

10.4.1Technical scheme of vedio based face recogntion in ATM165

10.4.2Flow chart of technical scheme166

10.4.3Introduction of fusion algorithm of face recognition167

10.4.4Experimental validation167

10.5Experiment results169

Chapter 11 Face disguise recognition and its application172

11.1Introduction and analysis173

11.1.1Real demand of face disguise recognition173

11.1.2Related work175

11.1.3Analysis of technical difficulties176

11.1.4Scheme of disguise difinition and recognition177

11.2Face disguise recognition methods179

11.2.1User location179

11.2.2Disguise detection based on Haar classifier180

11.2.3Classification of false alarm 186

11.2.4Elimination of false alarm for background using linear detection188

11.2.5Judgement of false alarm for sunglasses using eyeglasses bridge detection189

11.2.6Recognition of sunglasses disguise based on symmetric difference191

11.2.7Judgement of false alarm for sunglasses based on Gabor filter192

11.2.8Judgement of false alarm for hat based on clustering195

11.2.9Recognition of mouth-muffle disguise based on skin color model200

11.2.10Filtering other false alarms201

11.3Implementation and evaluation of face disguise recognition system201

11.3.1System structure201

11.3.2Design of system modules202

Chapter 12Face attendance and recognitioin system 206

12.1Infrared face recognition and bimodal face attendance system206

12.1.1Comparsion Far-infrared face image with near-infrared face image207

12.1.2Design of system207

12.1.3Software and hardware of system 208

12.2Recognition system combining face and fingerprint212

12.2.1Analysis and design of system212

12.2.2Software and hardware of system214

12.2.3Details of system215

12.3Computer login system based on face image authentication217

12.3.1Introduction217

12.3.2Windows login system217

12.3.3Design of face login system218

12.3.4Face recognition220

References222

"