Introduction / 1: |
The Problem / 1.1: |
Applications / 1.1.1: |
The Computer Graphics Approach / 1.1.2: |
Avoiding the Model / 1.1.3: |
A Review of Stereo Vision / 1.2: |
Camera Model and Image Formation / 1.2.1: |
Stereo Geometry / 1.2.2: |
The Correspondence Problem / 1.2.3: |
The Epipolar Constraint / 1.2.4: |
A Simple Stereo Geometry / 1.2.5: |
Rectification / 1.2.6: |
Example: SSD / 1.2.7: |
Contributions and Outline / 1.3: |
A Survey of Image-Based Rendering and Stereo / 2: |
Image-Based Rendering / 2.1: |
View Synthesis Based on Stereo / 2.1.1: |
View Interpolation / 2.1.2: |
Mosaics and Layered Representations / 2.1.3: |
Stereo / 2.2: |
A Framework for Stereo / 2.2.1: |
Preprocessing / 2.2.2: |
Matching Cost / 2.2.3: |
Evidence Aggregation / 2.2.4: |
Disparity Selection / 2.2.5: |
Sub-Pixel Disparity Computation / 2.2.6: |
Diffusion-Based Techniques / 2.2.7: |
Other Techniques / 2.2.8: |
Promising Recent Approaches / 2.2.9: |
Computer Vision Books / 2.3: |
View Synthesis / 3: |
Geometry / 3.1: |
Three-View Rectification / 3.1.1: |
The Linear Warping Equation / 3.1.2: |
Computing the Rectifying Homographies / 3.1.3: |
Synthesizing a New View / 3.2: |
Resolving Visibility / 3.2.1: |
Holes and Sampling Gaps / 3.2.2: |
Combining Information from Both Images / 3.2.3: |
Adjusting Intensities / 3.2.4: |
Filling Holes / 3.2.5: |
The View Synthesis Algorithm / 3.2.6: |
Limitations of the Approach / 3.2.7: |
Experiments / 3.3: |
Image-Based Scene Representations / 3.4: |
Summary / 3.5: |
Re-evaluating Stereo / 4: |
Traditional Applications of Stereo / 4.1: |
Automated Cartography / 4.1.1: |
Robot Navigation / 4.1.2: |
3D Reconstruction / 4.1.3: |
3D Recognition / 4.1.4: |
Visual Servoing / 4.1.5: |
Full vs. Weak Calibration / 4.1.6: |
Comparison of Requirements / 4.1.7: |
Stereo for View Synthesis / 4.2: |
Accuracy / 4.3: |
Correct vs. Realistic Views / 4.4: |
Areas of Uniform Intensities / 4.5: |
Geometric Constraints / 4.5.1: |
Interpolated Views / 4.5.2: |
Extrapolated Views / 4.5.3: |
General Views and the Aperture Problem / 4.5.4: |
Assigning Canonical Depth Interpretations / 4.5.5: |
Does Adding More Cameras Help? / 4.5.6: |
Partial Occlusion / 4.6: |
Gradient-Based Stereo / 4.7: |
Similarity and Confidence / 5.1: |
Displacement-Oriented Stereo / 5.2: |
The Evidence Measure / 5.3: |
Comparing Two Gradient Vectors / 5.3.1: |
Comparing Gradient Fields / 5.3.2: |
Computing Gradients of Discrete Images / 5.3.3: |
Accumulating the Measure / 5.4: |
Stereo: 1D Search Range / 5.5: |
General Motion: 2D Search Range / 5.5.3: |
Computing Disparity Maps for View Synthesis / 5.6: |
Occlusion Boundaries / 5.6.1: |
Detecting Partially Occluded Points and Uniform Regions / 5.6.2: |
Extrapolating the Disparities / 5.6.3: |
Efficiency / 5.7: |
Discussion and Possible Extensions / 5.8: |
Stereo Using Diffusion / 5.9: |
Disparity Space / 6.1: |
The SSD Algorithm and Boundary Blurring / 6.2: |
Aggregating Support by Diffusion / 6.3: |
The Membrane Model / 6.3.1: |
Support Function for the Membrane Model / 6.3.2: |
Diffusion with Local Stopping / 6.4: |
A Bayesian Model of Stereo Matching / 6.5: |
The Prior Model / 6.5.1: |
The Measurement Model / 6.5.2: |
Explicit Local Distribution Model / 6.5.3: |
Conclusion / 6.6: |
Contributions in View Synthesis / 7.1: |
Contributions in Stereo / 7.2: |
Extensions and Future Work / 7.3: |
Bibliography |
Introduction / 1: |
The Problem / 1.1: |
Applications / 1.1.1: |
The Computer Graphics Approach / 1.1.2: |
Avoiding the Model / 1.1.3: |
A Review of Stereo Vision / 1.2: |