Bridge the Vision Gap from Field to Command: A Deep Learning Network Enhancing Illumination and Details

Low-light image enhancement aims to improve the brightness of images captured in low-light conditions, which has various applications in surveillance, remote sensing and computational photography. However, low-light images often suffer from poor visibility and blurring, and simply brightening the dark regions may amplify blurring and cause detail loss. In this paper, we propose a simple yet effective two-stream framework called NEID that can enhance the brightness and the details simultaneously without introducing much computational cost. Specifically, our method consists of three modules: Light Enhancement (LE), Detail Refinement (DR) and Feature Fusing (FF), that aggregates composite features for multiple tasks using a channel attention mechanism. Extensive experiments on several benchmark datasets show the effectiveness of our method and its superiority over state-of-the-art methods.