Stereo Matching

Naive stereo matching with sliding window and stereo matching with dynamic programming for standard stereo image pairs.

Project structure

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.

Name		Name	Last commit message	Last commit date
Latest commit History 22 Commits
data_modified		data_modified
disparity_maps		disparity_maps
disparity_maps_dp		disparity_maps_dp
points_3d		points_3d
results		results
README.md		README.md
StereoMatchingDynamicProgramming.ipynb		StereoMatchingDynamicProgramming.ipynb
StereoMatchingNaiveSlidingWindow.ipynb		StereoMatchingNaiveSlidingWindow.ipynb
metrics.py		metrics.py
sm_dynamic_programming.py		sm_dynamic_programming.py
sm_naive_sliding_window.py		sm_naive_sliding_window.py
stereo_reconstruction.py		stereo_reconstruction.py