Acknowledgements

I. Representative images and 1-2 training/testing images for demonstration have been taken from:

"Stereo Vision: Algorithms and Applications" by Stefano Mattocia
Fischer et al., FlowNet: Learning Optical Flow with Convolutional Networks. ICCV 2015
Mayer et al., A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation. CVPR (2015)
Luo et al., Efficient Deep Learning for Stereo Matching. CVPR 2016
Scene Flow: https://lmb.informatik.uni-freiburg.de/resources/datasets/SceneFlowDatasets.en.html
KITTI 2015: http://www.cvlibs.net/datasets/kitti/eval_scene_flow.php?benchmark=stereo
Middlebury: http://vision.middlebury.edu/stereo
ETH3D: https://www.eth3d.net/low_res_two_view

The file has been slightly modified to suite our execution environment

Make sure you are inside "Session_3" directory and then open "session_3_depth_hands_on.ipynb" notebook
Execute the initial 4-5 lines in the notebook: they are for cloning the git repository (if required) and merging and moving weight files
- NOTE: Github doesn't allow file of size more than 100 MB, so I had to split each weight file into two parts
main.py, model.py and dataloader.py constitute the final merged code, refer them to know how to write a full working code after following the jupyter notebook hands-on. Run following command to learn about required arguments in main.py file:

python main.py -h

python main.py --help