Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Fang, Wenxuan, Weng, Jiangwei, Qian, Jianjun, Yang, Jian, Li, Jun
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2509.23150
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916974127218688
author Fang, Wenxuan
Weng, Jiangwei
Qian, Jianjun
Yang, Jian
Li, Jun
author_facet Fang, Wenxuan
Weng, Jiangwei
Qian, Jianjun
Yang, Jian
Li, Jun
contents Unsupervised image restoration under multi-weather conditions remains a fundamental yet underexplored challenge. While existing methods often rely on task-specific physical priors, their narrow focus limits scalability and generalization to diverse real-world weather scenarios. In this work, we propose \textbf{WeatherCycle}, a unified unpaired framework that reformulates weather restoration as a bidirectional degradation-content translation cycle, guided by degradation-aware curriculum regularization. At its core, WeatherCycle employs a \textit{lumina-chroma decomposition} strategy to decouple degradation from content without modeling complex weather, enabling domain conversion between degraded and clean images. To model diverse and complex degradations, we propose a \textit{Lumina Degradation Guidance Module} (LDGM), which learns luminance degradation priors from a degraded image pool and injects them into clean images via frequency-domain amplitude modulation, enabling controllable and realistic degradation modeling. Additionally, we incorporate a \textit{Difficulty-Aware Contrastive Regularization (DACR)} module that identifies hard samples via a CLIP-based classifier and enforces contrastive alignment between hard samples and restored features to enhance semantic consistency and robustness. Extensive experiments across serve multi-weather datasets, demonstrate that our method achieves state-of-the-art performance among unsupervised approaches, with strong generalization to complex weather degradations.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23150
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle WeatherCycle: Unpaired Multi-Weather Restoration via Color Space Decoupled Cycle Learning
Fang, Wenxuan
Weng, Jiangwei
Qian, Jianjun
Yang, Jian
Li, Jun
Computer Vision and Pattern Recognition
Unsupervised image restoration under multi-weather conditions remains a fundamental yet underexplored challenge. While existing methods often rely on task-specific physical priors, their narrow focus limits scalability and generalization to diverse real-world weather scenarios. In this work, we propose \textbf{WeatherCycle}, a unified unpaired framework that reformulates weather restoration as a bidirectional degradation-content translation cycle, guided by degradation-aware curriculum regularization. At its core, WeatherCycle employs a \textit{lumina-chroma decomposition} strategy to decouple degradation from content without modeling complex weather, enabling domain conversion between degraded and clean images. To model diverse and complex degradations, we propose a \textit{Lumina Degradation Guidance Module} (LDGM), which learns luminance degradation priors from a degraded image pool and injects them into clean images via frequency-domain amplitude modulation, enabling controllable and realistic degradation modeling. Additionally, we incorporate a \textit{Difficulty-Aware Contrastive Regularization (DACR)} module that identifies hard samples via a CLIP-based classifier and enforces contrastive alignment between hard samples and restored features to enhance semantic consistency and robustness. Extensive experiments across serve multi-weather datasets, demonstrate that our method achieves state-of-the-art performance among unsupervised approaches, with strong generalization to complex weather degradations.
title WeatherCycle: Unpaired Multi-Weather Restoration via Color Space Decoupled Cycle Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.23150