Abstract:Low-light images often suffer from severe noise, low brightness, low contrast, and color deviation. While several low-light image enhancement methods have been proposed, there remains a lack of efficient methods that can simultaneously solve all of these problems. In this paper, we introduce FLW-Net, a Fast and LightWeight Network for low-light image enhancement that significantly improves processing speed and overall effect. To achieve efficient low-light image enhancement, we recognize the challenges of the lack of an absolute reference and the need for a large receptive field to obtain global contrast. Therefore, we propose an efficient global feature information extraction component and design loss functions based on relative information to overcome these challenges. Finally, we conduct comparative experiments to demonstrate the effectiveness of the proposed method, and the results confirm that FLW-Net can significantly reduce the complexity of supervised low-light image enhancement networks while improving processing effect. Code is available at https://github.com/hitzhangyu/FLW-Net