An implementation of "Towards Loss-Resilient Image Coding for Unstable Satellite Networks" (AAAI 2025 Oral).
Our method builds on the channel-wise progressive coding framework, incorporating Spatial-Channel Rearrangement (SCR) on the encoder side and Mask Conditional Aggregation (MCA) on the decoder side to improve reconstruction quality with unpredictable errors. By integrating the Gilbert-Elliot model into the training process, we enhance the model’s ability to generalize in real-world network conditions.
If you use this code in your research, please cite our paper:
@inproceedings{sha2025towards,
title={Towards Loss-Resilient Image Coding for Unstable Satellite Networks},
author={Sha, Hongwei and Dong, Muchen and Luo, Quanyou and Lu, Ming and Chen, Hao and Ma, Zhan},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={39},
number={12},
pages={12506--12514},
year={2025}
}
-
Clone the repo
git clone https://github.com/NJUVISION/LossResilientLIC.git cd LossResilientLIC
-
Install dependencies
pip install compressai pip install timm
python train.py train_config.yaml
python eval_channel_packet.py
python progressive_test.py
├── bin/ # Compressed bit stream
├── exp/ # Organized network data
├── losses/ # RD loss
├── models/ # Model definitions
├── rec_img/ # Decompressed images
├── sim2net/ # Simple Network Simulator
├── train.py # Training entrypoint
├── eval_channel_packet.py # Evaluation entrypoint with packet
├── codec.py # Real inference Codec
├── train_config.yaml # Training config
├── codec_config.yaml # Real inference Codec config
├── progressive_test.py # Progressive inference entrypoint
└── README.md
BSD 3-Clause Clear License
This work builds upon the CompressAI, sim2net and ProgDTD.