RMFAT: Recurrent Multi-scale Feature Atmospheric Turbulence Mitigator

Zhiming Liu, Nantheera Anantrasirichai · 2025

Read RMFAT: Recurrent Multi-scale Feature Atmospheric Turbulence Mitigator by Zhiming Liu, Nantheera Anantrasirichai (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-2508-11409.txt.

Description

Zhiming Liu and Nantheera Anantrasirichai clean up video blurred by hot, shaky air in a 2025 paper. Heat haze warps shapes, softens detail, and makes frames flicker, which hurts cameras that watch roads, borders, and far scenes. Past top tools read many frames at once with heavy models, too slow and fat for live use on small gear.

Their tool, RMFAT, fixes each frame from just a pair of inputs in a light loop. Traits are read at many scales, and warp steps at both ends keep time flow smooth. On made up and real haze sets it beat past tools on clarity by near nine percent. It also ran more than four times faster.

Catalog record

CREATORZhiming Liu, Nantheera Anantrasirichai
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-11409

Browse the full library · About the Tower of Babel