L^2UWE: A Framework for the Efficient Enhancement of Low-Light\n Underwater Images Using Local Contrast and Multi-Scale Fusion

Abstract

Images captured underwater often suffer from suboptimal illumination settings\nthat can hide important visual features, reducing their quality. We present a\nnovel single-image low-light underwater image enhancer, L^2UWE, that builds on\nour observation that an efficient model of atmospheric lighting can be derived\nfrom local contrast information. We create two distinct models and generate two\nenhanced images from them: one that highlights finer details, the other focused\non darkness removal. A multi-scale fusion process is employed to combine these\nimages while emphasizing regions of higher luminance, saliency and local\ncontrast. We demonstrate the performance of L^2UWE by using seven metrics to\ntest it against seven state-of-the-art enhancement methods specific to\nunderwater and low-light scenes. Code available at:\nhttps://github.com/tunai/l2uwe.\n