An Image to Tailor: I-Frame Domain Adaptation in Neural Video Compression

Abstract

Neural video compression (NVC) models recently outperformed traditional methods. They typically include an I-Frame codec for Intra-Frames and a P-Frame codec for P-frames. However, their performance may be far from optimal with data outside the training set. We propose domain adaptation (DA) in NVC using lightweight convolutional adapters inserted in the I-Frame decoder of a pre-trained NVC model, which are then fine-tuned. These adapters shift knowledge to a specific domain without altering the architecture or causing catastrophic forgetting. They enhance compression for both I-frames and P-frames while using minimal parameters with respect to the entire architecture, improving NVC robustness.

Cite

Text

Presta et al. "An Image to Tailor: I-Frame Domain Adaptation in Neural Video Compression." NeurIPS 2024 Workshops: Compression, 2024.

Markdown

[Presta et al. "An Image to Tailor: I-Frame Domain Adaptation in Neural Video Compression." NeurIPS 2024 Workshops: Compression, 2024.](https://mlanthology.org/neuripsw/2024/presta2024neuripsw-image/)

BibTeX

@inproceedings{presta2024neuripsw-image,
  title     = {{An Image to Tailor: I-Frame Domain Adaptation in Neural Video Compression}},
  author    = {Presta, Alberto and Spadaro, Gabriele and Fiandrotti, Attilio and Grangetto, Marco},
  booktitle = {NeurIPS 2024 Workshops: Compression},
  year      = {2024},
  url       = {https://mlanthology.org/neuripsw/2024/presta2024neuripsw-image/}
}