HPGN: Hybrid Priors-Guided Network for Compressed Low-Light Image Enhancement

2025-04-03

AI Summary

To address the coupled degradation of insufficient illumination and JPEG artifacts in compressed low-light images—which severely compromises enhancement quality—this paper proposes the first unified joint enhancement framework. Our method innovatively fuses the JPEG quality factor and the DCT quantization matrix into a hybrid prior, enabling a plug-and-play module that achieves robust enhancement across multiple compression levels with a single model. The framework comprises four key components: DCT-domain feature modeling, a hybrid-prior-guided network, a stochastic quality-factor training strategy, and collaborative optimization along dual paths—compression and illumination. Extensive experiments on multiple benchmark datasets demonstrate significant improvements over state-of-the-art methods, with average gains of 1.27 dB in PSNR and 0.023 in SSIM. Moreover, our approach exhibits strong generalization capability and efficient inference speed.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

In practical applications, conventional methods generate large volumes of low-light images that require compression for efficient storage and transmission. However, most existing methods either disregard the removal of potential compression artifacts during the enhancement process or fail to establish a unified framework for joint task enhancement of images with varying compression qualities. To solve this problem, we propose the hybrid priors-guided network (HPGN), which enhances compressed low-light images by integrating both compression and illumination priors. Our approach fully utilizes the JPEG quality factor (QF) and DCT quantization matrix (QM) to guide the design of efficient joint task plug-and-play modules. Additionally, we employ a random QF generation strategy to guide model training, enabling a single model to enhance images across different compression levels. Experimental results confirm the superiority of our proposed method.

Problem

• Enhancing compressed low-light images effectively
• Integrating compression and illumination priors jointly
• Handling varying compression levels with one model

Innovation

• Integrates compression and illumination priors
• Uses JPEG QF and DCT QM
• Random QF strategy for training