Abstract:Exposure correction is a fundamental problem in computer vision and image processing. Recently, frequency domain-based methods have achieved impressive improvement, yet they still struggle with complex real-world scenarios under extreme exposure conditions. This is due to the local convolutional receptive fields failing to model long-range dependencies in the spectrum, and the non-generative learning paradigm being inadequate for retrieving lost details from severely degraded regions. In this paper, we propose Omnidirectional Spectral Mamba (OSMamba), a novel exposure correction network that incorporates the advantages of state space models and generative diffusion models to address these limitations. Specifically, OSMamba introduces an omnidirectional spectral scanning mechanism that adapts Mamba to the frequency domain to capture comprehensive long-range dependencies in both the amplitude and phase spectra of deep image features, hence enhancing illumination correction and structure recovery. Furthermore, we develop a dual-domain prior generator that learns from well-exposed images to generate a degradation-free diffusion prior containing correct information about severely under- and over-exposed regions for better detail restoration. Extensive experiments on multiple-exposure and mixed-exposure datasets demonstrate that the proposed OSMamba achieves state-of-the-art performance both quantitatively and qualitatively.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to perform image exposure correction under extreme exposure conditions. Specifically, although existing frequency - domain - based methods have made significant progress in dealing with exposure correction, in complex real - world scenarios, especially under extreme exposure conditions, there are still the following two main problems: 1. **Challenges in Illumination Correction and Structure Restoration**: Due to the limitations of the local convolutional receptive field, existing methods are difficult to capture long - distance dependencies in the frequency spectrum, which leads to poor performance in illumination correction and structure restoration. 2. **Difficulties in Detail Restoration**: In severely under - exposed or over - exposed areas, high - frequency details are severely lost, and it is difficult for discriminative networks trained only with regression losses to restore these details. To solve these problems, the paper proposes **Omnidirectional Spectral Mamba (OSMamba)**, a new type of exposure correction network. The main innovations of OSMamba include: - **Omni - directional Spectral Scanning (OS - Scan)**: By introducing the omni - directional spectral scanning mechanism, OSMamba can capture comprehensive long - distance dependencies of the magnitude and phase spectra in the frequency domain, thereby enhancing the ability of illumination correction and structure restoration. - **Dual - Domain Prior Generator (DDPG)**: Utilizing the capabilities of the generative diffusion model, it learns to generate a non - degraded diffusion prior from well - exposed images, which contains correct information in severely under - exposed and over - exposed areas to better restore details. Through these innovations, OSMamba has achieved state - of - the - art performance on multiple exposure correction datasets, performing well in both quantitative metrics and qualitative results.

OSMamba: Omnidirectional Spectral Mamba with Dual-Domain Prior Generator for Exposure Correction

ECMamba: Consolidating Selective State Space Model with Retinex Guidance for Efficient Multiple Exposure Correction

FD-Vision Mamba for Endoscopic Exposure Correction

ExpoMamba: Exploiting Frequency SSM Blocks for Efficient and Effective Image Enhancement

Dual Illumination Estimation for Robust Exposure Correction

SSUMamba: Spatial-Spectral Selective State Space Model for Hyperspectral Image Denoising

CDMamba: Remote Sensing Image Change Detection with Mamba

Multi-Scale Exposure Fusion Via Content Adaptive Edge-Preserving Smoothing Pyramids.

Spatial-Mamba: Effective Visual State Space Models via Structure-Aware State Fusion

Holistic Dynamic Frequency Transformer for Image Fusion and Exposure Correction

An optimized denoised bias correction model with local pre-fitting function for weak boundary image segmentation

MambaIRv2: Attentive State Space Restoration

ACMamba: A State Space Model-Based Approach for Multi-Weather Degraded Image Restoration

ColorMamba: Towards High-quality NIR-to-RGB Spectral Translation with Mamba

Learning Multi-Scale Photo Exposure Correction

Exposure Normalization and Compensation for Multiple-Exposure Correction

Pan-Mamba: Effective pan-sharpening with State Space Model

ExposureDiffusion: Learning to Expose for Low-light Image Enhancement

Exposure Correction Model to Enhance Image Quality

O-Mamba: O-shape State-Space Model for Underwater Image Enhancement