Low-light image enhancement is a challenging problem in the field of computer vision. In order to obtain more pleasing enhancement results, a low-light image enhancement method via joint convolutional sparse representation is proposed. The method is based on the Retinex theory and improves the problem of insufficient constraints. More concretely, when estimating illumination, the joint convolution sparse representation is proposed as structure and texture constraints to obtain a structural image severed as illumination. Then, the adaptive gradient constraint is used to enhance the details of the reflection image. Experiments on a number of challenging low-light images are present to reveal the efficacy of our method and show its superiority over several state-of-the-arts on both subjective and objective assessments.
In this paper, a novel enhancement algorithm for low-light images captured under low illumination conditions is proposed. More concretely, we design a method firstly to synthesize low-light images as training datasets. Then preclustering is conducted to separate training data into several groups by a coupled Gaussian mixture model. For each group, we adopt a coupled dictionary learning approach to train the low-light and normal-light dictionary pair jointly, and the statistical dependency of the sparsity coefficients is captured via Extreme Learning Machine simultaneously. Besides, we use a multi-phase dictionary learning strategy to enhance the robustness of our method. Experimental results show that proposed method is superior to existing methods.