Naive stereo matching with sliding window and stereo matching with dynamic programming for standard stereo image pairs.
- data_modified - Contains the input images from https://vision.middlebury.edu/stereo/data/
- disparity_maps - Computed disparity maps with naive stereo algorithm.
- disparity_maps_dp - Computed disparity maps with dynamic programming. *points3d - Computed (x, y, z, r, g, b) coordinates for stereo reconstruction.
- results - Csv files contains the results of the experiments.
- metrics.py - Cost functions to measure image similarity.
The other files are self-explanatory. Jupyter notebooks are for demonstration with examples.