Beyond Illumination: A Conditional Mutual Information-Guided Network for Low-Light Image Enhancement
2026-08-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors focus on improving low-light image enhancement, which means making dark photos look clearer and more natural. They point out that previous methods treated color and brightness separately but missed how these two aspects should work together. To fix this, they created a new network called CMIG-Net that uses conditional mutual information to better connect color and brightness details, adjusting the color based on local light conditions. Their tests show this approach works better than earlier methods, especially on very dark images.
Low-light image enhancementChrominanceIntensityHVI color spaceConditional mutual informationDual-branch architecturePSNRImage restorationMutual information calibration
Authors
Ya-nan Guan, Shaonan Zhang, Tao Dai, Tianqu Zhuang, Yongchao Qiao, Zhensen Chen, Shu-Tao Xia, Hang Guo
Abstract
Low-light image enhancement (LLIE) seeks to restore structural fidelity, natural color rendition, and proper exposure from images captured under inadequate lighting conditions. Recent state-of-the-art approaches, such as CIDNet, adopt a dual-branch architecture comprising a chrominance (HV) branch and an intensity (I) branch to separately model decoupled chromatic and luminance information within the HVI color space. However, these methods overlook the mutual interaction between intensity and chrominance components, which inherently limits their representational capacity and leads to suboptimal enhancement performance. To address this limitation, we propose the Conditional Mutual Information-Guided Network (CMIG-Net), which leverages conditional mutual information as a principled metric to quantitatively assess the contribution of chrominance features conditioned on the available intensity information. In particular, we design a Conditional Mutual Information Calibration (CMIC) module that generates a conditional information map, enabling region-adaptive recalibration of chrominance representations according to local illumination statistics. Furthermore, we introduce a Dynamic Dual-branch Information Restoration (D2IR) module, which adaptively governs bidirectional information flow between the intensity and chrominance branches, guided by both the conditional prior and the instantaneous restoration state. Extensive experiments on paired LLIE benchmarks demonstrate that CMIG-Net consistently outperforms CIDNet, achieving up to a 0.619 dB gain in PSNR, with a 0.382 dB improvement specifically on the challenging Sony-Total-Dark dataset.