This repository contains our series of works on Deep Hierarchical Video Compression.
- DHVC 1.0: The first hierarchical predictive coding method moves away from the hybrid coding framework, achieving best-in-class performance. Paper is available at Deep Hierarchical Video Compression (AAAI 2024).
- DHVC 2.0: The enhanced hierarchical predictive coding method, which integrates variable-rate intra- and inter-coding into a single model, delivering not only superior compression performance to representative methods but also real-time processing with a significantly smaller memory footprint on standard GPUs. Paper is available at High-Efficiency Neural Video Compression via Hierarchical Predictive Learning (arxiv 2024).
- DHIC: A spectrally regularized hierarchical image coding method, delivering not only superior compression performance but also efficient optimization without increasing inference complexity. Paper is available at Taming Hierarchical Image Coding Optimization: A Spectral Regularization Perspective (ICLR 2026).
[2026.5.21] We have reconstructed the code and uploaded the pretrained models of DHIC.
[2026.1.26] Excited to share that our work on intra-frame hierarchical coding (DHIC) has been accepted by ICLR! The code will be open-sourced soon. 🚀
[2025.2.12] We have reconstructed the code and uploaded the pretrained models of DHVC 1.0.
If you find this work helpful to your research, please cite:
@inproceedings{lu2024deep,
title={Deep Hierarchical Video Compression},
author={Lu, Ming and Duan, Zhihao and Zhu, Fengqing and Ma, Zhan},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={38},
number={8},
pages={8859--8867},
year={2024}
}
@article{lu2024high,
title={High-Efficiency Neural Video Compression via Hierarchical Predictive Learning},
author={Lu, Ming and Duan, Zhihao and Cong, Wuyang and Ding, Dandan and Zhu, Fengqing and Ma, Zhan},
journal={arXiv preprint arXiv:2410.02598},
year={2024}
}
@inproceedings{congtaming,
title={Taming Hierarchical Image Coding Optimization: A Spectral Regularization Perspective},
author={Cong, Wuyang and Shi, Junqi and Lu, Ming and Zhang, Xu and Ma, Zhan},
booktitle={The Fourteenth International Conference on Learning Representations}
}
If you have any question, feel free to contact us via minglu@nju.edu.cn or congwuyang@smail.nju.edu.cn.