Improved water-net model for underwater image enhancement
Deepa Abin, Sudeep D. Thepade, Amulya R. Maitre · AIP conference proceedings · 2024
Exploring the unknown has always fascinated humans and life underwater is no exception.Hence, AUVs or ROVs are used to discover the unreachable depths of the water bodies.Since the environmental conditions differ from the air domain, the raw images/videos captured are blurred or have low contrast.The current proposal is to improve the previously introduced Water-Net CNN model for enriching underwater images.The objective is to enhance the visual perception of the resultant enhanced images by replacing the input of simple histogram equalization with a cascading fusion block of CLAHE and BPDFHE algorithms.The model is evaluated using PSNR and SSIM full-reference evaluation metrics.