Rethinking Model Redundancy for Low-light Image Enhancement

TL;DR

This paper addresses the issue of model redundancy in low-light image enhancement by proposing two techniques, ADR and POG, which improve performance by reducing harmful and useless parameters.

Contribution

It introduces novel methods, ADR and POG, to effectively mitigate model redundancy in LLIE models, leading to enhanced image enhancement performance.

Findings

Proposed techniques outperform baseline models in LLIE tasks.

Mitigating parameter redundancy improves image quality.

Code will be publicly released for reproducibility.

Abstract

Low-light image enhancement (LLIE) is a fundamental task in computational photography, aiming to improve illumination, reduce noise, and enhance the image quality of low-light images. While recent advancements primarily focus on customizing complex neural network models, we have observed significant redundancy in these models, limiting further performance improvement. In this paper, we investigate and rethink the model redundancy for LLIE, identifying parameter harmfulness and parameter uselessness. Inspired by the rethinking, we propose two innovative techniques to mitigate model redundancy while improving the LLIE performance: Attention Dynamic Reallocation (ADR) and Parameter Orthogonal Generation (POG). ADR dynamically reallocates appropriate attention based on original attention, thereby mitigating parameter harmfulness. POG learns orthogonal basis embeddings of parameters and prevents degradation to static parameters, thereby mitigating parameter uselessness. Experiments validate the effectiveness of our techniques. We will release the code to the public.