Fourier-RWKV: A Multi-State Perception Network for Efficient Image Dehazing

2025-12-08

AI Summary

Non-uniform haze in real-world scenarios causes severe image degradation, yet existing Transformer-based methods suffer from quadratic computational complexity, hindering real-time deployment. To address this, we propose a multi-state-aware linear-complexity dehazing network, introducing a novel paradigm that jointly models spatial, frequency-domain, and semantic information. Specifically, we extend the RWKV attention mechanism to the Fourier domain, design a deformable four-directional token shifting module for local adaptivity, incorporate a Fourier mixing block to capture long-range frequency-domain dependencies, and employ dynamic semantic kernels for cross-domain feature alignment. Our method achieves state-of-the-art performance across multiple benchmarks, reduces computational cost by 47–63%, and attains 52 FPS inference speed on 1080p images—marking the first end-to-end real-time dehazing framework that maintains high restoration fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Hardware-aware ML
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Security and Privacy: Data transparency and provenance

Abstract

Image dehazing is crucial for reliable visual perception, yet it remains highly challenging under real-world non-uniform haze conditions. Although Transformer-based methods excel at capturing global context, their quadratic computational complexity hinders real-time deployment. To address this, we propose Fourier Receptance Weighted Key Value (Fourier-RWKV), a novel dehazing framework based on a Multi-State Perception paradigm. The model achieves comprehensive haze degradation modeling with linear complexity by synergistically integrating three distinct perceptual states: (1) Spatial-form Perception, realized through the Deformable Quad-directional Token Shift (DQ-Shift) operation, which dynamically adjusts receptive fields to accommodate local haze variations; (2) Frequency-domain Perception, implemented within the Fourier Mix block, which extends the core WKV attention mechanism of RWKV from the spatial domain to the Fourier domain, preserving the long-range dependencies essential for global haze estimation while mitigating spatial attenuation; (3) Semantic-relation Perception, facilitated by the Semantic Bridge Module (SBM), which utilizes Dynamic Semantic Kernel Fusion (DSK-Fusion) to precisely align encoder-decoder features and suppress artifacts. Extensive experiments on multiple benchmarks demonstrate that Fourier-RWKV delivers state-of-the-art performance across diverse haze scenarios while significantly reducing computational overhead, establishing a favorable trade-off between restoration quality and practical efficiency. Code is available at: https://github.com/Dilizlr/Fourier-RWKV.

Problem

• Efficiently removes non-uniform haze from images
• Reduces computational complexity for real-time deployment
• Integrates spatial, frequency, and semantic perception for restoration

Innovation

• Linear complexity via Multi-State Perception paradigm
• Frequency-domain attention extends RWKV to Fourier domain
• Dynamic Semantic Kernel Fusion aligns encoder-decoder features