• MindSpore reproduction of DSIC
• • Our contributions
  • Files
  • Environment
  • Command
  • Reproduction results:
  • Cite from:
  • Contributors:

MindSpore reproduction of DSIC

original paper link

Paper Citation:
@inproceedings{liu2019dsic,
title={Dsic: Deep stereo image compression},
author={Liu, Jerry and Wang, Shenlong and Urtasun, Raquel},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages={3136--3145},
year={2019}
}

Our contributions

1. transplant from Pytorch to mindspore (including train,test scripts and model architecture).
2. benchmark tests on the origin datasets.
3. compare performance between different frameworks.

Files

to be updated

Environment

to be updated

Command

to be updated

Reproduction results:

to be updated

Cite from:

The authors do not release the source code. The original model architecture is integrated in https://github.com/ywz978020607/HESIC/tree/master/ywz/DSIC/. The train and test scripts used in this project are also referenced by HESIC.

Contributors:

name: Huiming Zheng (the School of Electronic and Computer Engineering, Shenzhen Graduate School, Peking University (PKU), Shenzhen, China)
email: hmzheng@stu.pku.edu.cn

If any problems about the project, please contact me with my email.