/CLIC-2021-Learned-Image-Compression

Primary LanguageJupyter Notebook

CLIC-2021-Learned-Image-Compression

This repository contains our solution to CLIC 2021

Method: ResHiFiC

ResHiFiC is a method based on Google HiFiC, it modifies the encoder architecture using ResNet

Usage

We provide a JupyterNotebook for you to train your model and compress your image. Run it by clicking "Open in Colab"

Files

  • compression: modified HiFiC model
  • Paper: papers we referenced
  • HiFiC_image: images compressed by original HiFiC model
  • ResHiFiC_image: images compressed by ResHiFiC (modified HiFiC) model
  • report_image: images used by final report
  • report: result reports
    • report.py: plotting script used by ResHiFiC.pdf
    • ResHiFiC.pdf: final report
    • 算分proj大报告.pptx: PPT for final demonstration
  • CLIC2021.ipynb: JupyterNotebook for training and testing out model, please open it in Google Colab