Skip to content
/ HFLIC Public

Officia code for Human Friendly Perceptual Learned Image Compression with Reinforced Transform and Unofficial Implementation of papar "PO-ELIC: Perception-Oriented Efficient Learned Image Coding."

Notifications You must be signed in to change notification settings

beiluo97/HFLIC

Repository files navigation

HFLIC

This section contains the official code for Human Friendly Perceptual Learned Image Compression with Reinforced Transform. Additionally, it includes an unofficial implementation of the paper titled "PO-ELIC: Perception-Oriented Efficient Learned Image Coding." The code implementation is based on CompressAI. We share our enhance transform elic ckpt in Enh-ELIC-ckpt, and modify the config_5group.py in ./config, you can train HFLIC and Enh-POELIC with different lambda. We further share some of our code and lamda setting for "EnhPO:ELIC" and "HFLIC"

PO:ELIC

This section consists of an unofficial implementation of the paper titled "PO-ELIC: Perception-Oriented Efficient Learned Image Coding," which was presented at CVPR22W as the 1st place winner of CLIC22.

How to train po:elic

To utilize the code effectively, follow these steps: Open the file modules/layers/res_blk.py and modify the ResidualBottleneck class. Set N * 2 to N / 2. Open the file config/config_5group.py and locate the line "lambda_face": 0. Modify this line to set the value of "lambda_face" as 0.

ELIC

We have incorporated the ELIC code from the GitHub repository maintained by JiangWeibeta. You can find the code and related resources at the following link: https://github.com/JiangWeibeta/ELIC.

Cite

HFLIC

@misc{ning2023hflic,
      title={HFLIC: Human Friendly Perceptual Learned Image Compression with Reinforced Transform}, 
      author={Peirong Ning and Wei Jiang and Ronggang Wang},
      year={2023},
      eprint={2305.07519},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

PO:ELIC

@inproceedings{he2022po,
  title={PO-ELIC: Perception-Oriented Efficient Learned Image Coding},
  author={He, Dailan and Yang, Ziming and Yu, Hongjiu and Xu, Tongda and Luo, Jixiang and Chen, Yuan and Gao, Chenjian and Shi, Xinjie and Qin, Hongwei and Wang, Yan},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={1764--1769},
  year={2022}
}

ELIC

@misc{jiang2022unofficialelic,
    author={Jiang, Wei},
    title={Unofficial ELIC},
    howpublished={\url{https://github.com/JiangWeibeta/ELIC}},
    year={2022}
}
@inproceedings{he2022elic,
  title={Elic: Efficient learned image compression with unevenly grouped space-channel contextual adaptive coding},
  author={He, Dailan and Yang, Ziming and Peng, Weikun and Ma, Rui and Qin, Hongwei and Wang, Yan},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={5718--5727},
  year={2022}
}

About

Officia code for Human Friendly Perceptual Learned Image Compression with Reinforced Transform and Unofficial Implementation of papar "PO-ELIC: Perception-Oriented Efficient Learned Image Coding."

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published