Single-Step Latent Diffusion for Underwater Image Restoration

Jiayi Wu, Tianfu Wang, Md Abu Bakr Siddique, Md Jahidul Islam, Cornelia Fermuller, Yiannis Aloimonos, Christopher A. Metzler · 2025

Read Single-Step Latent Diffusion for Underwater Image Restoration by Jiayi Wu, Tianfu Wang, Md Abu Bakr Siddique, Md Jahidul Islam, Cornelia Fermuller, Yiannis Aloimonos, Christopher A. Metzler (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-07878.txt.

Description

Jiayi Wu and a team of co-authors restore underwater photos in a single diffusion step instead of hundreds. Their work appeared in twenty twenty-five. Murk steals color and contrast, yet marine science, farming, and archaeology all depend on clear images.

Their network, SLURPP, pairs a pretrained latent diffusion model with an explicit split of the scene. That split lets the model account for light fading and scattering with depth, the failure mode of earlier methods on complex scenes. Training data came from a physics-based pipeline that drowned ordinary photos realistically, with full annotations. On synthetic tests the method ran over two hundred times faster than diffusion rivals and scored about three decibels cleaner. Real-world images improved convincingly too.

Catalog record

CREATORJiayi Wu, Tianfu Wang, Md Abu Bakr Siddique, Md Jahidul Islam, Cornelia Fermuller, Yiannis Aloimonos, Christopher A. Metzler
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-07878

Browse the full library · About the Tower of Babel