Low-Light Image Enhancement Using Zero-DCE and DCP

In this paper, a new low-light enhancement technique is proposed to enhance the performance of images. This technique is obtained by fusing zero-reference deep curve estimation (Zero-DCE) and dark channel prior (DCP). We calculate image-specific parameter curves using convolutional neural networks (CNNs), which enhance the low-light image pixel-wise. The proposed method follows a spatial attention mechanism to emphasize appropriate regions. This method is simple and effective for low-light enhancement. Further, the performance of the proposed method is validated and compared in terms of peak signal to noise ratio (PSNR), structural index similarity (SSIM) and run time (RT) with other existing methods. The qualitative and quantitative results show the superior performance of the proposed method compared to other existing methods.