Acquisition, Compression and Rendering of Depth and Texture for Multi-View Video

Yannick Morvan

June 9, 2009



Chapter1 Introduction
 1.1 Fundamentals of stereo visualization
 1.2 Applications of multi-view imaging
 1.3 Three-dimensional video systems layout
 1.4 Multi-view acquisition, compression and rendering problems addressed in this thesis
 1.5 Thesis outline: the multi-view video system
 1.6 Contributions of this thesis

Chapter2 Projective Geometry
 2.1 Projective geometry
 2.2 Pinhole camera model
 2.3 Camera calibration
 2.4 Two-view geometry
 2.5 Summary and conclusions

Chapter3 Multi-View Depth Estimation
 3.1 Introduction
 3.2 Two-view depth estimation
 3.3 Previous work on depth estimation
 3.4 Multiple-view depth estimation
 3.5 Summary and conclusions

Chapter4 Multi-View Depth Image Based Rendering
 4.1 Introduction
 4.2 Depth Image Based Rendering
 4.3 Occlusion-compatible scanning order
 4.4 Occlusion handling
 4.5 Experimental results on rendering quality and evaluation
 4.6 Conclusions

Chapter5 H.264-Based Depth and Texture Multi-View Coding
 5.1 Introduction
 5.2 Multi-view video coding tools
 5.3 View Synthesis Prediction (VSP) for N-depth/N-texture coding
 5.4 Experimental results
 5.5 Conclusions

Chapter6 Depth image coding using piecewise-linear functions
 6.1 Introduction
 6.2 Geometric modeling of depth images
 6.3 Bit-allocation strategy
 6.4 Experimental results
 6.5 Conclusions and perspectives

Chapter7 Joint Depth-Texture Bit Allocation
 7.1 Introduction
 7.2 Joint depth/texture bit allocation
 7.3 Hierarchical search optimization
 7.4 Relationship between the depth and texture bit rate
 7.5 Applications of the joint bit-allocation framework
 7.6 Experimental results
 7.7 Conclusions

Chapter8 Conclusions and Prospects
 8.1 Results and discussion on the individual chapters
 8.2 Key issues and open questions
 8.3 Perspectives

Chapter9 Appendix
 9.1 Appendix. Occlusion-compatible scanning-order
 9.2 Appendix. Test sequences and images