Skip to content
main
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
Apr 21, 2021
Apr 21, 2021
Apr 21, 2021
Apr 21, 2021
Apr 21, 2021
Apr 21, 2021

Scalable Near-lossless Image Compression

Source code of our CVPR 2021 paper "Learning Scalable ℓ-constrained Near-lossless Image Compression via Joint Lossy Image and Residual Compression".

Usage

To run the code requires Python 3.6 and TensorFlow 1.15.

pip install tensorflow-compression==1.3
pip install range-coder

The main.py provides compress and decompress functions, and an example to encode/decode ./test_patch/kodim05_p128.png. Users can test their own images.

The model ckp_003 can be downloaded from Baidu Netdisk, with access code snic.

Please note: the current implementation is not optimized for speed. Residual compression is slow. We are working on the fast version 💻.

Citation

@article{Bai_2021_SNIC,
  title={Learning Scalable $\ell_\infty$-constrained Near-lossless Image Compression via Joint Lossy Image and Residual Compression},
  author={Bai, Yuanchao and Liu, Xianming and Zuo, Wangmeng and Wang, Yaowei and Ji, Xiangyang},
  journal={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2021}
}

About

Source code of CVPR 2021 paper "Learning Scalable ℓ∞-constrained Near-lossless Image Compression via Joint Lossy Image and Residual Compression".

Resources

Releases

No releases published

Packages

No packages published

Languages