What question did this study set out to answer?

This research aims to enhance low-light images using a novel adaptive framework that leverages semantic priors.

April 16, 2026

Adaptive Semantic Priors Guided Multiple Scale Low-Light Image Enhancement

Key Points

This research aims to enhance low-light images using a novel adaptive framework that leverages semantic priors.
Developed an Adaptive Semantic Prior Guided Multi-Scale Low-Light Image Enhancement framework.
Utilized a U-Net backbone with a lightweight semantic prior extractor.
Implemented a dynamic illumination integration strategy for improved results.
Achieved 23.8 dB PSNR on the LOL test set and 0.86 SSIM at 0.18 seconds inference time.
Demonstrated 88.4% pixel accuracy on VE-LOL, outperforming existing methods such as Retinex-Net and EnlightenGAN.

Abstract

To address those major weaknesses of current methods of low-light enhancement, we introduce an Adaptive Semantic Prior Guided Multi-Scale Low-Light Image Enhancement (ASPG-MSLIE) framework which consists of three sections as illustrated in the abstract. In order to resolve this problem we suggest a U-Net backbone with a lightweight semantic prior extractor and a dynamic illumination integration strategy. The network is capable of maintaining the perceptual relevance of high-level semantic context, and suppressing noise in other semantic contexts, through conditioning improvement. So-dimensionality is confirmed by experiments on LOL test set (23.8 dB PSNR, 0.86 SSIM at 0.18 s inferring time) and VE-LOL (88.4% pixel accuracy), confirming state-of-the-art performance over Retinex-Net, KinD, Zero-DCE, and EnlightenGAN.

Bookmark

Cite This Study

Harish et al. (Thu,) studied this question.

synapsesocial.com/papers/69e07cfa2f7e8953b7cbe096 https://doi.org/https://doi.org/10.1051/itmconf/20268501011/pdf

Bookmark