This is the implementation of our paper"SBIN: A stereo disparity estimation network using binary convolutions"

• docker
• nvidia-docker
• Download Kitti2012, Kitti2015 and Sceneflow datasets
  • Datasets must be inside the folder datasets in the root folder of this repository
  • Datasets
    • kitti2012
    • kitti2015
    • sceneflow

Log and checkpoints are created on the folder output

./docker/launch.sh

# Sceneflow training
python experiments.py train 2

# Kitti2012 training 
python experiments.py train 1

To add new experiments, check the experiments folder of directly use the cli.py, after reading the command line help

@article{Aguilera_2022,
title={SBIN: A stereo disparity estimation network using binary convolutions},
volume={20},
url={https://latamt.ieeer9.org/index.php/transactions/article/view/5909},
number={4}, journal={IEEE Latin America Transactions},
author={Aguilera, Cristhian Alejandro},
year={2022},
month={Jan.}, 
pages={693–699} }