Boosting Fidelity for Pre-Trained-Diffusion-Based Low-Light Image Enhancement via Condition Refinement

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Xiaogang, Wang, Jian, Lu, Yunfan, Chu, Ruihang, Wang, Ruixing, Wu, Jiafei, Yu, Bei, Lin, Liang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915563401379840
author Xu, Xiaogang
Wang, Jian
Lu, Yunfan
Chu, Ruihang
Wang, Ruixing
Wu, Jiafei
Yu, Bei
Lin, Liang
author_facet Xu, Xiaogang
Wang, Jian
Lu, Yunfan
Chu, Ruihang
Wang, Ruixing
Wu, Jiafei
Yu, Bei
Lin, Liang
contents Diffusion-based methods, leveraging pre-trained large models like Stable Diffusion via ControlNet, have achieved remarkable performance in several low-level vision tasks. However, Pre-Trained Diffusion-Based (PTDB) methods often sacrifice content fidelity to attain higher perceptual realism. This issue is exacerbated in low-light scenarios, where severely degraded information caused by the darkness limits effective control. We identify two primary causes of fidelity loss: the absence of suitable conditional latent modeling and the lack of bidirectional interaction between the conditional latent and noisy latent in the diffusion process. To address this, we propose a novel optimization strategy for conditioning in pre-trained diffusion models, enhancing fidelity while preserving realism and aesthetics. Our method introduces a mechanism to recover spatial details lost during VAE encoding, i.e., a latent refinement pipeline incorporating generative priors. Additionally, the refined latent condition interacts dynamically with the noisy latent, leading to improved restoration performance. Our approach is plug-and-play, seamlessly integrating into existing diffusion networks to provide more effective control. Extensive experiments demonstrate significant fidelity improvements in PTDB methods.
format Preprint
id arxiv_https___arxiv_org_abs_2510_17105
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Boosting Fidelity for Pre-Trained-Diffusion-Based Low-Light Image Enhancement via Condition Refinement
Xu, Xiaogang
Wang, Jian
Lu, Yunfan
Chu, Ruihang
Wang, Ruixing
Wu, Jiafei
Yu, Bei
Lin, Liang
Computer Vision and Pattern Recognition
Diffusion-based methods, leveraging pre-trained large models like Stable Diffusion via ControlNet, have achieved remarkable performance in several low-level vision tasks. However, Pre-Trained Diffusion-Based (PTDB) methods often sacrifice content fidelity to attain higher perceptual realism. This issue is exacerbated in low-light scenarios, where severely degraded information caused by the darkness limits effective control. We identify two primary causes of fidelity loss: the absence of suitable conditional latent modeling and the lack of bidirectional interaction between the conditional latent and noisy latent in the diffusion process. To address this, we propose a novel optimization strategy for conditioning in pre-trained diffusion models, enhancing fidelity while preserving realism and aesthetics. Our method introduces a mechanism to recover spatial details lost during VAE encoding, i.e., a latent refinement pipeline incorporating generative priors. Additionally, the refined latent condition interacts dynamically with the noisy latent, leading to improved restoration performance. Our approach is plug-and-play, seamlessly integrating into existing diffusion networks to provide more effective control. Extensive experiments demonstrate significant fidelity improvements in PTDB methods.
title Boosting Fidelity for Pre-Trained-Diffusion-Based Low-Light Image Enhancement via Condition Refinement
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.17105