Video Understanding

This project aims to investigate and propose methods to improve the generalization ability of deep learning in video understanding tasks, aiming to reduce the need for human supervision.