Comparative Study of Low-Light Image Enhancement Performance: DCE-Net vs. MirNetv2

The difficult computer vision task of low-light image augmentation tries to raise the standard of photographs taken in low light. Deep learning-based low-light picture enhancing systems have recently attained modern outcomes. In this article, we compare the effectiveness of two cutting-edge deep learning-based algorithms-DCE-Net and MirNetv2-to enhance the quality of images in low light. A deep cross-enhancement network called DCE-Net uses a cutting-edge cross-connection module to take advantage of the complementing data that exists between the network's various tiers. A multi-scale feature extraction module is used by the multi-scale illumination recovery network MirNetv2 to extract features from various scales of the picture. On the LOLdataset benchmark Datasets for enhancing low-light images, we compare the execution of DCE-Net and MirNetv2. The outcomes demonstrate that DCE-Net and MirNetv2 provides cutting-edge outcomes. On the LOLdataset, DCE-Net performs better than MirNetv2 in terms of improving contrast and brightness. The findings of this study indicate that the deep learning-based Improving visibility in low-light images method DCE-Net- a promising one. In a wide range of applications, including surveillance, night photography, and medical imaging, DCE-Net can be utilized to improve low-light photos.