tallamjr / clic

Challenge for Learning Image Compression, 2021

Home Page:http://compression.cc/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

CVPR Workshop: Challenge for Learning Image Compression

2020 Papers

Evaluation Metrics

The evaluate() in metrics.py function expects two dictionaries mapping filenames fo filepaths (the target PNGs and PNGs produced by the decoder).

I attached the code that's used to evaluate submissions. The function evaluate() in metrics.py receives two dictionaries containing paths to the target PNGs and decoded PNGs. In the video challenge, these are the PNGs representing Y, U, and V channels. One thing to note is that the MS-SSIM is weighted by the number of pixels, so Y is weighted more strongly than U or V.

  • Lucas

About

Challenge for Learning Image Compression, 2021

http://compression.cc/

License:MIT License


Languages

Language:Python 100.0%