Home

Awesome

Context-adaptive Entropy Model for End-to-end Optimized Image Compression

Repository of the paper "Context-adaptive Entropy Model for End-to-end Optimized Image Compression"

Introduction

This repository includes evaluation results, reconstructed images, and test codes of our paper "Context-adaptive Entropy Model for End-to-end Optimized Image Compression". Please refer to the paper for the detailed information. If the paper or this repository helps you, please cite our work as:

@InProceedings{Lee2019Context,
    author = {Lee, Jooyoung and Cho, Seunghyun and Beack, Seung-Kwon},
    title = {Context-adaptive Entropy Model for End-to-end Optimized Image Compression},
    booktitle = {the 7th Int. Conf. on Learning Representations},
    month = {May},
    year = {2019}
}

[Updated on Mar 31, 2020] We've uploaded the second version of our test codes. The second version uses THE SAME MODELS as for the first version. If you've already downloaded our models before, just copy them into the "models" directory under the working directory. The followings are the new features of the updated test codes.

Samples
Test results over the CLIC validation set
Samples
Test results over the Tecnick SAMPLING imageset

Reconstructed samples

Samples

Evaluation results

We optimized the networks with the two different types of distortion terms, one with MSE and the other with MS-SSIM. For each distortion type, the average bits per pixel (BPP) and the distortion, PSNR or MS-SSIM, over 24 PNG images of the Kodak PhotoCD image dataset are measured for each of the nine R-D configurations. Therefore, a total of 18 networks are trained and evaluated. To our best knowledge, this is the first test results of ANN-based image compression which outperform BPG, in terms of PSNR as well as MS-SSIM.

Followings are the rate-distortion curves of the proposed method and competitive methods. The top plot represents PSNR values in consequence of bpp changes, while the bottom plot shows MS-SSIM values in the same manner. Note that MS-SSIM values are converted to decibels for differentiating the quality levels.

RD-PSNR

RD-MS-SSIM

The compression gains in terms of BD-rate of PSNR over JPEG2000, Balle'18(MSE-optimized), BPG are 34.08%, 11.87%, 6.85%, respectively. In case of MS-SSIM, we found the wider gaps of 68.82%, 13.93%, 49.68%, respectively.

Test code [DEPRECATED]

[Updated on Mar 31, 2020] Please use the second version of our test codes for more efficient and exact evaluation. Its instructions are provided via readme.text in the second version of our test codes. The following instructions for the previous version of our test code will be deleted from our repositary sooner or later.

We provide test codes for evaluating a image compression model proposed in our work "Context-adaptive Entropy Model for End-to-end Optimized Image Compression". The model includes two types of models, base models for low bpp and hybrid models for high bpp.

Supported I/O formats

Python Imaging Library (PIL) is used in our test code. To file the supported formats, please refer to https://pillow.readthedocs.io/en/4.1.x/handbook/image-file-formats.html. PNG and BMP formats including RGB channels have been tested.

Encoding

Decoding

Notes