Multiple weather degraded image restoration based on multi-component decomposition

doi:10.1016/j.patcog.2025.112561

Multiple weather degraded image restoration based on multi-component decomposition

2025-10-08

0

PRE

AI

DOI：10.1016/j.patcog.2025.112561

Original

Original request for help

Save

Abstract

En 中文

• Three-component prior decomposes images into background, primary mask, and residual. • Transformer trunk and attention branches disentangle overlapping weather degradations. • Unified MCD-Net restores single-weather and multi-weather images with outstanding performance.

Multiple weather degraded image restoration based on multi-component decomposition

Abstract

Journal

Organization