Stereo Matching using Spatial Search Trees
Partially complete implementation of stereo matching for rectified image pairs using octree-like representations.
Tasks (to complete sometime in the future):
- tree representation
- naive matching using similarity
- support for image/videos
- homogeneity of regions
- piece-wise continuity
- vertical smoothness of disparity
- multi-objective optimization function (genetic algorithms)
- tree encoding for space optimization
- utilizing tree as local context for misc tasks
Good References:
- Fast Stereo Matching Using Rectangular Subregioning and 3D Maximum-Surface Techniques, C Sun, IJCV, 2002. pdf
- Multi-Resolution Rectangular Subregioning Stereo Matching Using Fast Correlation and Dynamic Programming Techniques, C Sun, 1998. pdf
- Fast Stereo Matching by Iterated Dynamic Programming and Quadtree Subregioning, C Leung, BMVC, 2004. pdf
- Stereo Matching Using Global and Local Segmentation, Kim et al. pdf
- Multi-resolution stereo matching using genetic algorithm, Gong et al, SMBV 2001. pdf
- Genetic-Based Stereo Algorithm and Disparity Map Evaluation, Gong et al, IJCV 2002. pdf
- Quadtree-based genetic algorithm and its applications to computer vision, Gong et al, 2004. pdf
- Octree Representations of Moving Objects, Ahuja et al, 1984. pdf
- Low-level Multiscale Image Segmentation and A Benchmark For Its Evaluation, Akbas et al, 2020. pdf