July 19, 2024Pattern RecognitionOpen Access

Denoising diffusion post-processing for low-light image enhancement

View Full Paper

Ask AI

Bookmark

Authors

SPSavvas PanagiotouUniversity of Pretoria ABAnna Sergeevna BosmanUniversity of Pretoria

Discussion

Member takes

Implication

Experimental evaluation demonstrates improved perceptual quality in enhanced low-light images, highlighting the utility of single-pass diffusion post-processing.

Key Points

Enhanced low-light images achieve higher perceptual quality without oversmoothing, outperforming conventional post-processing denoisers across challenging low-light datasets.
Benchmark testing across challenging low-light datasets validates a single-pass diffusion model, modeling the conditional distribution between under-exposed and normal images.
Avoiding the computationally expensive generative reverse process enables single-pass execution, reducing image degradation and color bias from low-light image enhancement.

Cite This Study

Panagiotou et al. (2024) studied this question.

synapsesocial.com/papers/68e5fa6bb6db64358758eed0 https://doi.org/10.1016/j.patcog.2024.110799

View Full Paper

Ask AI

Bookmark

Implication

Experimental evaluation demonstrates improved perceptual quality in enhanced low-light images, highlighting the utility of single-pass diffusion post-processing.

Denoising diffusion post-processing for low-light image enhancement

Discussion

Member takes

Implication

Key Points

Cite This Study

Denoising diffusion post-processing for low-light image enhancement

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider