| Low Level Approach for Image and Video Understanding | |
| GOP Structure Conversion in Transcoding MPEG-2 to H.264/AVC | p. 3 |
| Introduction | p. 3 |
| GOP Structure Conversion | p. 5 |
| MV Scaling in the Temporal Direction | p. 5 |
| Correlation Between the Current MB Mode and the Reference Region | p. 6 |
| Proposed Algorithms | p. 8 |
| Adaptive Search Range Selection through the MV Linearity Test in the Temporal Direction | p. 8 |
| Adaptive Mode Decision Method Based on Region Information | p. 9 |
| Simulation Results | p. 13 |
| Conclusion | p. 16 |
| References | p. 16 |
| Simple Low Level Features for Image Analysis | p. 17 |
| Introduction | p. 17 |
| The Role of Color | p. 19 |
| Color Spaces | p. 20 |
| HSL and HSV | p. 23 |
| CIE-Lab | p. 25 |
| Color Flattening | p. 26 |
| Blob Detection | p. 27 |
| Edge Detection | p. 30 |
| Simple Shapes | p. 33 |
| Scale and Position Invariants: Procrustes Analysis | p. 33 |
| Shape Alignment: Iterative Closest Point | p. 34 |
| Shape Encoding and Matching: Curvature Space Scale | p. 35 |
| Combination of simple features | p. 37 |
| Conclusions | p. 38 |
| References | p. 39 |
| Fast and robust Face Detection | p. 43 |
| Related work | p. 43 |
| Feature-based | p. 44 |
| Low level | p. 44 |
| Skin-map | p. 45 |
| Feature analysis | p. 48 |
| Template based | p. 49 |
| Appearance-based | p. 50 |
| Introduction | p. 52 |
| Face detection on video stream | p. 52 |
| Efficient object detection | p. 53 |
| Appearance-based face detection | p. 56 |
| Features-based face detection | p. 58 |
| Adaptive Skin detection | p. 59 |
| Eyes detection and validation | p. 61 |
| Face normalization | p. 64 |
| Experimental results | p. 64 |
| Discussion and Conclusions | p. 68 |
| References | p. 69 |
| Automatic 3D Facial Fitting for Tracking in Video Sequence | p. 73 |
| The 3D Face Model | p. 73 |
| 3D morphing basis | p. 75 |
| 3D morphing basis for shape and expression | p. 77 |
| Shape Unit | p. 77 |
| Expression Unit | p. 79 |
| Appearance basis | p. 81 |
| PCA | p. 81 |
| Image-based PCA | p. 82 |
| 3D Illumination basis | p. 87 |
| The General Purposes 3D Tracking Algorithm | p. 90 |
| Feature Location | p. 92 |
| Model adaptation | p. 94 |
| Feature-based pose estimation | p. 95 |
| Shape and expression inference | p. 98 |
| 3D Tracking-based Model refinement | p. 101 |
| Initial refinement | p. 101 |
| Deep refinement | p. 103 |
| Experimental results | p. 105 |
| References | p. 110 |
| Multimedia Knowledge-Based Approaches and Applications | |
| Input Devices and Interaction Techniques for VR-Enhanced Medicine | p. 115 |
| Introduction | p. 116 |
| Related Works | p. 117 |
| Requirements Analysis | p. 119 |
| Interaction Metaphors and Techniques | p. 121 |
| Realistic Metaphors | p. 122 |
| A Realistic Metaphor: Virtual Hand | p. 122 |
| Magic Metaphors | p. 122 |
| A Magic Metaphor: Virtual Pointer | p. 123 |
| Pros and Cons of Realistic vs. Magic Interaction Metaphors | p. 123 |
| The Proposed Input Device: the Wiimote | p. 124 |
| Communication | p. 125 |
| Inputs | p. 125 |
| Outputs | p. 125 |
| Classification | p. 125 |
| The Proposed Interaction Techniques | p. 126 |
| The Manipulation State | p. 126 |
| Pointing | p. 127 |
| Translation and Zooming | p. 127 |
| Rotation | p. 128 |
| The Cropping State | p. 130 |
| Discussion | p. 132 |
| References | p. 132 |
| Bridging Sensing and Decision Making in Ambient Intelligence Environments | p. 135 |
| Introduction | p. 136 |
| Related Works | p. 137 |
| Preliminaries | p. 139 |
| Templates | p. 141 |
| Uncertainty Resolver via Aggregation Functions | p. 142 |
| Average-based Function | p. 143 |
| Bayesian Network-Based Function | p. 143 |
| "Dempster and Shafer" -Based Function | p. 146 |
| Decision Tree-Based Function | p. 148 |
| Experimentation | p. 150 |
| Aggregation Function Accuracy and Time Processing | p. 151 |
| Value Distribution | p. 153 |
| Test 1: Values higher than 0.5 | p. 153 |
| Test 2: Values less than 0.5 | p. 154 |
| Test 3: Random Values | p. 154 |
| Test 5: 75% of the values are less than 0.5 | p. 156 |
| Test 6: Equally distributed values | p. 157 |
| Test 7: Distribution change | p. 158 |
| Test 8: Influence of the number of returned values 0 and 1 on the aggregated result | p. 158 |
| Discussion | p. 159 |
| Template Tuning | p. 160 |
| Case 1: using the multimedia function f1 | p. 161 |
| Case 2: using the multimedia function f2 | p. 162 |
| Uncertainty threshold tuning | p. 162 |
| Conclusion | p. 163 |
| References | p. 163 |
| Ambient Intelligence in Multimedia and Virtual Reality Environments for the rehabilitation | p. 165 |
| Introduction | p. 166 |
| Using AI by special needs users | p. 167 |
| Visual Impairment and Partially Sighted People | p. 167 |
| Deaf and Hard-of-Hearing People | p. 168 |
| Physically Disabled Persons | p. 168 |
| Mentally Disabled People | p. 169 |
| Smart Home | p. 169 |
| A detailed example of using AI in virtual reality for rehabilitation | p. 170 |
| Future vision | p. 174 |
| Conclusion | p. 175 |
| Acknowledgement | p. 175 |
| References | p. 175 |
| Artificial Neural Networks for Processing Graphs with Application to Image Understanding: A Survey | p. 179 |
| From flat to structural Pattern Recognition | p. 179 |
| Graph processing by neural networks | p. 183 |
| Notation | p. 183 |
| A general framework for graph processing | p. 184 |
| Recursive Neural Networks | p. 186 |
| Graph Neural Networks | p. 187 |
| Other models | p. 188 |
| Graph-based representation of images | p. 189 |
| Image segmentation | p. 189 |
| Region Adjacency Graphs | p. 191 |
| Multi-resolution trees | p. 194 |
| Conclusions | p. 197 |
| References | p. 197 |
| Index | p. 201 |
| Table of Contents provided by Ingram. All Rights Reserved. |
