Global-Local Collaborative Enhancement: A Low-Light Image Enhancement Framework Guided by Adaptive Adjustment

Images captured under low-light conditions often suffer from issues such as insufficient brightness and color distortion, severely limiting the advancement of image observation technology. To address these challenges, we propose a Global- Local Collaborative Enhancement (GLCE) network that progressively enhances low-light inputs, ensuring improved visibility and color fidelity. First, we use a joint adaptive enhancement module to extract image features, which dynamically adjusts them to optimize global brightness distribution and achieve balanced illumination. Then, in the second stage, multi-scale convolution blocks are applied to enhance illumination and reflectance components separately, capturing details at different scales and improving information representation in dark regions. Experiments on two datasets demonstrate that our method achieves significant contrast enhancement and color restoration advantages.