Low-light image enhancement based on the combination of adaptive variational model and alternating least square method

Existing low-light enhancement methods often suffer from inadequate enhancement in dark regions or over-exposure and color distortion in well-lit areas due to significant brightness variations in real-world scenes. Furthermore, ambient light interference frequently causes color inaccuracies in enhanced images. To address these issues, we propose a low-light image enhancement method based on an improved adaptive variational model. First, building upon existing models, we implement adaptive region partitioning and introduce a global quadratic dispersion term. Finally, local brightness adjustment and global color correction are optimized via the alternating least-squares algorithm. This paper establishes the existence and uniqueness of the minimizer for the variational minimization problem. The Split Bregman technique is employed to achieve efficient numerical implementation of the adaptive variational model. Additionally, numerical experiments are conducted to compare the proposed method with other classical enhancement techniques, demonstrating its effectiveness.