WeatherSeg: Weather-Robust Image Segmentation using Teacher-Student Dual Learning and Classifier-Updating Attention

arXiv cs.CV / 4/28/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

WeatherSeg is a semi-supervised semantic segmentation framework designed for autonomous driving perception under adverse weather such as rain, clouds, and fog, while lowering required annotation effort.
The method uses a Dual Teacher-Student Weight-Sharing Model (DTSWSM) to distill knowledge from weather-degraded images, improving robustness beyond standard training.
WeatherSeg also introduces Classifier Weight Updating Attention (CWUAM), which dynamically updates classifier weights according to environmental attributes to better handle changing conditions.
Experiments report that WeatherSeg outperforms baseline segmentation models in both accuracy and weather robustness across multiple conditions, supporting its use for all-weather applications.
The arXiv release positions WeatherSeg as a practical research direction for improving perception reliability in real-world, weather-varying driving scenarios.

Abstract

WeatherSeg, an advanced semi-supervised segmentation framework, addresses autonomous driving's environmental perception challenges in adverse weather while reducing annotation costs. This framework integrates a Dual Teacher-Student Weight-Sharing Model (DTSWSM) that enables knowledge distillation from weather-affected images, and a Classifier Weight Updating Attention Mechanism (CWUAM) that dynamically adjusts classifier weights based on environmental attributes. Comprehensive evaluations demonstrate that WeatherSeg significantly outperforms baseline models in both accuracy and robustness across various weather conditions, including clear, rainy, cloudy, and foggy scenarios, establishing it as an effective solution for all-weather semantic segmentation in autonomous driving and related applications.