Learning from Failure: Leveraging Unreliable Predictions in Semi-Supervised Real-World Adverse Weather Removal

2026-10-01

AI Summary

This study addresses the limited real-world generalization and lack of semantic constraints in existing adverse weather image restoration models, which predominantly rely on synthetic data supervision. To overcome these limitations, this work proposes a semi-supervised contrastive learning framework based on a student-teacher architecture. Notably, it pioneers the utilization of unreliable predictions as negative samples to enhance the robustness of contrastive learning. Furthermore, an efficient phase spectrum semantic prior is introduced to replace text-based supervision, accompanied by a newly designed adaptive phase consistency loss function. Extensive evaluations on real-world benchmarks demonstrate that the proposed method surpasses current state-of-the-art models in both restoration quality and perceptual fidelity, significantly improving generalization capability against authentic adverse weather degradations.

Abstract

Adverse weather image restoration aims to recover images degraded by rain, haze, snow, and other weather-induced artifacts, thereby improving the robustness of outdoor vision systems. Existing unified restoration models exhibit limited generalization to real-world scenes due to their reliance on synthetic supervision and insufficient semantic constraints. In this paper, we propose a novel student--teacher semi-supervised framework that addresses both challenges. Specifically, we introduce an unreliable database that preserves failed teacher predictions as informative negative samples for contrastive learning, while a reliable database stores high-quality teacher predictions as positive samples. By jointly exploiting reliable pseudo-ground truths and unreliable teacher outputs, the proposed framework learns to enhance desirable restoration characteristics while avoiding common failures. We further propose a phase spectrum-based semantic constraint that replaces computationally expensive text-based supervision with an efficient and naturally aligned semantic prior. An adaptive phase consistency loss is also designed to dynamically balance supervision between the degraded input and teacher pseudo-ground truths according to degradation severity. Extensive experiments on real-world benchmarks demonstrate that the proposed method consistently outperforms existing state-of-the-art approaches in restoration quality and perceptual fidelity while exhibiting stronger generalization to real-world adverse weather conditions.

Problem

• adverse weather removal
• image restoration
• semi-supervised learning
• generalization
• semantic constraint

Innovation

• Semi-supervised learning
• Contrastive learning
• Phase spectrum constraint
• Adverse weather removal
• Student-teacher framework