Dynamic Low-Rank Instance Adaptation for Universal Neural Image Compression

TL;DR

This paper introduces a dynamic low-rank adaptation method for neural image compression that effectively reduces domain gap issues, improves rate-distortion performance on out-of-domain images, and maintains low bit rate overhead.

Contribution

The paper proposes a novel dynamic low-rank adaptation approach with a gating network, enhancing universal neural image compression across diverse datasets.

Findings

Achieves approximately 19% BD-rate improvement on out-of-domain images.

Outperforms existing instance adaptive methods by about 5% BD-rate.

Effectively mitigates domain gap in neural image compression.

Abstract

The latest advancements in neural image compression show great potential in surpassing the rate-distortion performance of conventional standard codecs. Nevertheless, there exists an indelible domain gap between the datasets utilized for training (i.e., natural images) and those utilized for inference (e.g., artistic images). Our proposal involves a low-rank adaptation approach aimed at addressing the rate-distortion drop observed in out-of-domain datasets. Specifically, we perform low-rank matrix decomposition to update certain adaptation parameters of the client's decoder. These updated parameters, along with image latents, are encoded into a bitstream and transmitted to the decoder in practical scenarios. Due to the low-rank constraint imposed on the adaptation parameters, the resulting bit rate overhead is small. Furthermore, the bit rate allocation of low-rank adaptation is \emph{non-trivial}, considering the diverse inputs require varying adaptation bitstreams. We thus introduce a dynamic gating network on top of the low-rank adaptation method, in order to decide which decoder layer should employ adaptation. The dynamic adaptation network is optimized end-to-end using rate-distortion loss. Our proposed method exhibits universality across diverse image datasets. Extensive results demonstrate that this paradigm significantly mitigates the domain gap, surpassing non-adaptive methods with an average BD-rate improvement of approximately $19\%$ across out-of-domain images. Furthermore, it outperforms the most advanced instance adaptive methods by roughly $5\%$ BD-rate. Ablation studies confirm our method's ability to universally enhance various image compression architectures.