Computer Vision and Action Recognition: A Guide for Image Processing and Computer Vision Community for Action Understanding

GetFullPageImage......Page 1
front-matter......Page 2
Computer Vision and Action Recognition......Page 4
Preface......Page 7
Foreword......Page 9
Acknowledgments......Page 11
Contents......Page 13
List of Figures......Page 18
List of Tables......Page 20
1.2 What is Action?......Page 21
1.3 Action Recognition in Computer Vision......Page 22
1.4 Application Realms of Action Recognition......Page 23
1.5 Categorization......Page 24
1.6 Think Ahead!......Page 28
2.2 Pre-processing Steps......Page 29
2.3 Segmentation and Extraction......Page 30
2.3.3 Corner Points......Page 31
2.3.4 Blob Detectors......Page 32
2.3.5 Feature Descriptors......Page 33
2.3.6 Segmentation......Page 34
2.3.6.1 Background Subtraction......Page 35
2.3.6.2 Frame Subtraction......Page 36
2.3.6.3 DenseMotion Estimation......Page 37
2.4 Local Binary Pattern......Page 40
2.4.1 LBP—Variants......Page 41
2.5 Structure fromMotion (SFM)......Page 43
2.5.1 Constraints of FM......Page 45
2.5.2 Improvements of FM......Page 47
2.6.1 Intensity Normalization......Page 54
2.6.2.2 Feature-based ImageMatching......Page 55
2.7 Think Ahead!......Page 56
3.2 Classification of Various Dimensions of Representations......Page 58
3.2.1 Bag-of-Features (BoF) or Bag-of-Visual-Words (BoVW)......Page 59
3.2.1.1 Constraints of BoF......Page 62
3.2.2 ProductManifold Approaches......Page 63
3.3.1 Interest-point-based Approaches......Page 68
3.3.2.3 Factorial HiddenMarkovModels......Page 72
3.3.2.7 Others......Page 73
3.3.3 Eigenspace-based Approach......Page 74
3.3.5.1 Actom SequenceModel......Page 75
3.3.5.2 Spatio-temporal Silhouettes......Page 76
3.3.5.3 Robustness in CameraMovement......Page 79
3.4 View-invariantMethods......Page 80
3.5 Gesture Recognition and Analysis......Page 81
3.7 Affective Computing and Expression Analysis......Page 82
3.7.1 Games with Emotional Involvement......Page 83
3.7.3 Anatomically-based Talking Head......Page 85
3.8.2 BasicMotion......Page 86
3.8.6 Direction-based BasicMotion......Page 87
3.9 Gait Analysis......Page 88
3.10.1 Application Areas......Page 90
3.10.2 RelatedWorks on Low-Resolution Video Processing......Page 91
3.11 Discussion......Page 92
3.11.1 Salient Region and Its Associated Salient Region Construction......Page 93
3.13 Think Ahead!......Page 94
4.1 Motion History Image (MHI)......Page 96
4.3.1 Formation of an MHI Image......Page 98
4.3.2 Motion Energy Image (MEI)......Page 99
4.3.5 Temporal Duration vs. Decay Parameter......Page 101
4.3.6.4 Outlier Removal......Page 104
4.4.1 Self-occlusion Problem......Page 105
4.4.9 Non-trajectory Nature......Page 107
4.5.2.1 Gait Energy Image (GEI)......Page 110
4.5.2.5 Dominant Energy Image (DEI)......Page 111
4.5.2.8 SuperposedMotion Image (SMI)......Page 112
4.5.2.12 Edge-based MHI......Page 113
4.6.1.1 Multiple-levelMHIMethod......Page 114
4.4.7 Not View-invariantMethod......Page 106
4.6.1.2 Directional MHIMethod......Page 115
4.6.1.3 HierarchicalMotion History HistogramMethod......Page 118
4.6.2 Solving Variable-lengthMovements......Page 119
4.6.4 Hierarchical-MHI......Page 120
4.6.8 Contour-based STV......Page 121
4.6.9.2 3DMotion HistoryModel......Page 122
4.6.9.6 Volumetric MHI......Page 123
4.8 Implementations of the MHI......Page 124
4.8.2 The MHI and its Variants in Analysis......Page 125
4.8.3 The MHI and its Variants in Interactions......Page 126
4.9 MHI and its Future......Page 130
4.10 Conclusion......Page 131
4.11 Think Ahead!......Page 132
4.2 Why MHI?......Page 97
4.3.6 Update Functionψ(x, y, t )......Page 102
4.3.6.2 Frame to Frame Difference......Page 103
5.1.1 Two-frame-based Approaches......Page 133
5.1.3 Discussions......Page 134
5.2 Shape Representation Schemes......Page 137
5.2.1.2 Global Approaches......Page 139
5.3.1 HuMoments for Feature Sets......Page 140
5.3.1.1 Low-OrderMoments......Page 141
5.3.2 ZernikeMoments for Feature Sets......Page 144
5.3.2.2 HuMoment—Some Points......Page 145
5.4.1 Appropriate Feature Selection......Page 146
5.4.2 Dimension Reduction......Page 147
5.4.2.4 Multifactor Kernel Principal Component Analysis (MKPCA)......Page 148
5.4.2.9 Linear Discriminant Analysis (LDA)......Page 149
5.4.2.14 Eigenspace......Page 150
5.5 Pattern Classifications......Page 151
5.5.1 Supervised Learning......Page 152
5.5.1.2 Boosting:......Page 153
5.5.2 Unsupervised Learning......Page 155
5.5.3 Nearest Neighbor......Page 156
5.5.3.1 Choice of k......Page 157
5.5.4 Cross-validation—Partitioning Scheme......Page 160
5.6 EvaluationMatrices......Page 161
5.7 Think Ahead!......Page 163
6.2 Necessity for Standard Datasets......Page 164
6.3.1 KTH Dataset......Page 166
6.3.2 Weizmann Dataset......Page 167
6.3.3 IXMAS Dataset......Page 168
6.3.7 Korea University Gesture Database......Page 169
6.4.2 Naval Air Training and Operating Procedures Standardization (NATOPS)Dataset......Page 170
6.5.1 Youtube Dataset......Page 171
6.5.4 UCF Sports Dataset......Page 172
6.5.8 Kisses/Slaps Dataset......Page 173
6.6 Datasets on Other Arenas......Page 174
6.6.3 Collective Activity Dataset......Page 175
6.6.5 People PlayingMusical Instrument (PPMI)......Page 178
6.6.10 HumanEva Dataset......Page 179
6.6.12 Other Datasets......Page 181
6.7 Challenges Ahead on Datasets......Page 182
6.8 Think Ahead!......Page 185
6.2.1 Motion Capture System......Page 165
7.2 Key Challenges Ahead in Action Recognition......Page 190
7.3 Few Points for New Researchers......Page 194
7.3.1 Intelligent Transport System......Page 197
7.5 Think Ahead!......Page 199
Bibliography......Page 200