BVI-Mamba: Video Enhancement Using a Visual State-Space Model for Low-Light and Underwater Environments

Videos captured in low-light and underwater conditions often suffer from distortions such as noise, low contrast, color imbalance, and blur. These issues not only limit visibility but also degrade automatic tasks like detection. Post-processing is typically required but can be time-consuming. AI-based tools for video enhancement also demand significantly more computational resources compared to image-based methods. This paper introduces a novel framework, Visual Mamba, designed to reduce memory usage and computational time by leveraging the Visual State Space (VSS) model. The framework consists of two modules: (i) a feature alignment module, where spatio-temporal displacement between input frames is registered in the feature space, and (ii) an enhancement module, where noise removal and brightness adjustment are performed using a UNet-like architecture, with all convolutional layers replaced by VSS blocks. Experimental results show that the Visual Mamba technique outperforms Transformer and convolution-based models in both low-light and underwater video enhancement tasks. Code is available on line at https://github.com/russellllaputa/BVI-Mamba.

TempRetinex: Retinex-based Unsupervised Enhancement for Low-light Video Under Diverse Lighting Conditions

Videos inherently contain rich temporal information that provides complementary cues for low-light enhancement beyond what can be achieved with single images. We propose TempRetinex, a novel unsupervised Retinex-based framework that effectively exploits inter-frame correlations for video enhancement. To address the poor generalization of existing unsupervised methods under varying illumination, we introduce adaptive brightness adjustment (ABA) preprocessing that explicitly aligns lighting distributions across exposures. This significantly improves model robustness to diverse lighting scenarios and eases training optimization, leading to better denoising performance. For enhanced temporal coherence, we propose a multi-scale temporal consistency-aware loss to enforce multiscale similarity between consecutive frames, and an occlusion-aware masking technique to handle complex motions. We further incorporate a reverse inference strategy to refine unconverged frames and a self-ensemble (SE) mechanism to boost the denoising across diverse textures. Experiments demonstrate that TempRetinex achieves state-of-the-art performance in both perceptual quality and temporal consistency, achieving up to a 29.7% PSNR gain over prior methods.

Zero-TIG: Temporal Consistency-Aware Zero-Shot Illumination-Guided Low-light Video Enhancement

Low-light and underwater videos suffer from poor visibility, low contrast, and high noise, necessitating enhancements in visual quality. However, existing approaches typically rely on paired ground truth, which limits their practicality and often fails to maintain temporal consistency. To overcome these obstacles, this paper introduces a novel zero-shot learning approach named Zero-TIG, leveraging the Retinex theory and optical flow techniques. The proposed network consists of an enhancement module and a temporal feedback module. The enhancement module comprises three subnetworks: low-light image denoising, illumination estimation, and reflection denoising. The temporal enhancement module ensures temporal consistency by incorporating histogram equalization, optical flow computation, and image warping to align the enhanced previous frame with the current frame, thereby maintaining continuity. Additionally, we address color distortion in underwater data by adaptively balancing RGB channels. The experimental results demonstrate that our method achieves low-light video enhancement without the need for paired training data, making it a promising and applicable method for real-world scenario enhancement.