An online, interactive, and collaborative platform for underwater video understanding

Abstract

This thesis introduces an advanced video analysis platform designed explicitly for underwater videos. The study focuses on developing a scalable software architecture that can handle the complexities of underwater footage analysis. The platform is enhanced with state-of-the-art machine learning models, enabling it to interpret data from underwater recordings effectively. This combination of elastic architecture and customized machine learning models offers a reliable solution for studying aquatic footage, supporting marine research. In addition, the platform employs a microservices architecture, augmented with Docker container support, to ensure scalability, streamlined operations, and ease of deployment. This design facilitates independent, flexible service management, boosting operational efficiency and adaptability. Overall, the thesis presents a solution that addresses the challenges of underwater video analysis, contributing substantially to the field of marine environment research and exploration by Computer Vision techniques.

Date of Award	2023
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Sai Kit YEUNG (Supervisor)