Skip to content

mahaichuan/iWave

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

iWave

Wavelet-like Transform with CNN re-implementation.

This repo provides the official implementation of "iWave: CNN-Based Wavelet-Like Transform for Image Compression".

Accepted by IEEE TMM.

Author: Haichuan Ma, Dong Liu, Ruiqin Xiong, Feng Wu

BibTeX

@article{ma2019iwave,

title={iWave: CNN-Based Wavelet-Like Transform for Image Compression},

author={Ma, Haichuan and Liu, Dong and Xiong, Ruiqin and Wu, Feng},

journal={IEEE Transactions on Multimedia},

year={2019},

publisher={IEEE}

}

Update Notes

2020.10.29 Upload code and pre-trained models.

2020.10.29 Init this repo.

How To Test

  1. Dependencies. We test with MIT deepo docker image.

  2. Transform an image with transform.py. PTAHs or image SIZE need to be modified.

  3. Reconstruct an image from iWave cofficients with inverse_transform.py. PTAHs or image SIZE need to be modified.

  4. Have fun!

Results

  1. Decompose image with iWave, compared with CDF9/7. (top: CDF9/7, bottom: iWave)

image

  1. Reconstructions (gray-scale image patches from Kodak-05 and Kodak-08), compared with JPEG-2000 (Jasper), at the same bit rate. (top: CDF9/7, bottom: iWave)

image

image

About

Wavelet-like Transform with CNN

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages