Low-light image enhancement method based on joint guiding total variation

Low-light image enhancement is crucial for high-quality image display and various visual applications. However, maintaining the visual naturalness of an image while enhancing its brightness presents a challenging task. We propose a novel algorithm for low-light image enhancement that joint guiding total variation and luminance adaptive correction. In order to obtain illuminated images with clear structure and smooth detail texture, the total variation of the window incorporating the guiding filter is used as the illumination regularity term. Logarithmic mean local variance is used as a reflection regularization term to extract texture details better reflecting the image. In addition, to address the issue of color distortion in areas with extremely low illumination, an arc-tangent transform is used to enhance the color and contrast in the very low-light region. We solve the model using an alternating optimization algorithm. In the model, each sub-problem is transformed into a vectorized least squares regression, with closed-form solutions. Experimental results on several public datasets demonstrate that our method suppresses the noise well while maintaining the color contrast and detail information. In comparison to other advanced algorithms, the proposed model delivers superior results in both subjective and objective evaluations.