Traffic Image Restoration under Adverse Weather via Frequency-Aware Mamba

TL;DR

This paper introduces FAMamba, a novel frequency-aware architecture for traffic image restoration under adverse weather, combining frequency-guided feature extraction and wavelet-based detail refinement to improve image quality.

Contribution

It proposes a new framework integrating frequency guidance into Mamba architecture, including dual-branch feature extraction and wavelet-based residual learning for enhanced restoration.

Findings

FAMamba outperforms existing methods in restoring degraded traffic images.

The frequency-aware components improve texture detail preservation.

Experiments validate the efficiency and effectiveness of the proposed approach.

Abstract

Traffic image restoration under adverse weather conditions remains a critical challenge for intelligent transportation systems. Existing methods primarily focus on spatial-domain modeling but neglect frequency-domain priors. Although the emerging Mamba architecture excels at long-range dependency modeling through patch-wise correlation analysis, its potential for frequency-domain feature extraction remains unexplored. To address this, we propose Frequency-Aware Mamba (FAMamba), a novel framework that integrates frequency guidance with sequence modeling for efficient image restoration. Our architecture consists of two key components: (1) a Dual-Branch Feature Extraction Block (DFEB) that enhances local-global interaction via bidirectional 2D frequency-adaptive scanning, dynamically adjusting traversal paths based on sub-band texture distributions; and (2) a Prior-Guided Block (PGB) that refines texture details through wavelet-based high-frequency residual learning, enabling high-quality image reconstruction with precise details. Meanwhile, we design a novel Adaptive Frequency Scanning Mechanism (AFSM) for the Mamba architecture, which enables the Mamba to achieve frequency-domain scanning across distinct subgraphs, thereby fully leveraging the texture distribution characteristics inherent in subgraph structures. Extensive experiments demonstrate the efficiency and effectiveness of FAMamba.