Guest Editorial: Introduction to the Special Issue on Advanced Machine Learning Methodologies for Underwater Image and Video Processing and Analysis

Abstract

In the realm of ocean engineering, underwater images and videos serve as vital carriers of information. However, the challenging conditions of underwater imaging often lead to quality degradation in captured content. These degradations, encompassing issues, such as diminished contrast, color casts, blurred details, and uneven brightness, not only hinder human perception but also present formidable obstacles for leveraging underwater media in ocean engineering applications. Despite advancements in the processing and analysis of underwater images and videos, the methodologies employed thus far have proven to be less than optimal. Furthermore, the direct application of established in-air techniques to underwater scenarios remains problematic due to the distinct attributes of underwater imaging, notably the effects of light selective absorption and scattering. As a result, there is a pressing need for fresh theories, methodologies, and applications that cater specifically to the challenges of processing and analyzing underwater visual content. Recent progress in advanced machine learning methodologies provides an avenue of promise, offering novel insights and approaches to address the issues of underwater images and videos.