DANICE: Domain Adaptation Without Forgetting in Neural Image Compression

Sudeep Katakol, Luis Herranz, Fei Yang, Marta Mrak; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2021, pp. 1921-1925

Abstract


Neural image compression (NIC) is a new coding paradigm where coding capabilities are captured by deep models learned from data. This data-driven nature enables new potential functionalities. In this paper, we study the adaptability of codecs to custom domains of interest. We show that NIC codecs are transferable and that they can be adapted with relatively few target domain images. However, naive adaptation interferes with the solution optimized for the original source domain, resulting in forgetting the original coding capabilities in that domain, and may even break the compatibility with previously encoded bitstreams. Addressing these problems, we propose Codec Adaptation without Forgetting (CAwF), a framework that can avoid these problems by adding a small amount of custom parameters, where the source codec remains embedded and unchanged during adaptation process. Experiments demonstrate its effectiveness and provide useful insights on the characteristics of catastrophic interference in NIC.

Related Material


[pdf] [supp] [arXiv]
[bibtex]
@InProceedings{Katakol_2021_CVPR, author = {Katakol, Sudeep and Herranz, Luis and Yang, Fei and Mrak, Marta}, title = {DANICE: Domain Adaptation Without Forgetting in Neural Image Compression}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2021}, pages = {1921-1925} }