Domain Adaptation for Learned Image Compression with Supervised Adapters
This work addresses domain adaptation for image compression, enabling more efficient compression across diverse image types, though it is incremental as it builds on existing pre-trained models.
The paper tackles the problem of adapting learned image compression models to multiple target domains without losing performance on the source domain, achieving improved rate-distortion efficiency on target domains and better encoding for out-of-domain images.
In Learned Image Compression (LIC), a model is trained at encoding and decoding images sampled from a source domain, often outperforming traditional codecs on natural images; yet its performance may be far from optimal on images sampled from different domains. In this work, we tackle the problem of adapting a pre-trained model to multiple target domains by plugging into the decoder an adapter module for each of them, including the source one. Each adapter improves the decoder performance on a specific domain, without the model forgetting about the images seen at training time. A gate network computes the weights to optimally blend the contributions from the adapters when the bitstream is decoded. We experimentally validate our method over two state-of-the-art pre-trained models, observing improved rate-distortion efficiency on the target domains without penalties on the source domain. Furthermore, the gate's ability to find similarities with the learned target domains enables better encoding efficiency also for images outside them.