提出光照-色彩解耦的新视角，提升暗光图像增强效果
Rethinking Low-Light Image Enhancement: A Log-Domain Intensity--Chromaticity Decoupling Perspective
- 从对数域解耦光照与色彩信息，避免异常放大和色噪
- 在LOLv2-Real上达到29.71dB PSNR和0.89 SSIM
- 改善低光下人脸检测，适合图像增强与下游任务应用
基于解耦表示显式施加重建约束，抑制异常通道放大和色噪声。在LOLv2-Real、MIT-Adobe FiveK和LSRW数据集上的实验表明，该方法在定量和视觉表现上均具竞争力或更优，在LOLv2-Real上达到29.71 dB PSNR和0.89 SSIM。DarkFace实验进一步证明其提升了低光条件下的下游人脸检测性能。代码与预训练模型已公开于：https://github.com/mubaisam/ICD。
原文摘要 · Abstract (English)
Explicit reconstruction constraints derived from the decoupled representation are further imposed to suppress abnormal channel amplification and chromatic noise. Experiments on LOLv2-Real, MIT-Adobe FiveK, and LSRW show that the proposed method achieves competitive or superior quantitative and visual performance, reaching 29.71 dB PSNR and 0.89 SSIM on LOLv2-Real. DarkFace experiments further indicate improved downstream face detection under low-light conditions. Code and pretrained models are available at: https://github.com/mubaisam/ICD.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。