Human motion is deceivingly simple but intricately complicated. We walk, run and jump on a daily basis without realising how much time and effort we made to master these motor skills when we were little. Understanding human motions is understanding humans ourselves.
The applications of human motion analysis broadly range from computer graphics (e.g. CG in movies and games), computer vision (e.g. activity recognition), to even health and medical (e.g. motor control, Parkinson disease) and sports (e.g. training athletes), etc.
This project aims to use the cutting-edge machine learning/deep learning method to analyse human motions, which can be further used for synthesising new natural motions for entertainment/visualisation purposes, understanding interactions between people in dancing and sports, understanding interactions between people and their immediate physical environments and predicting/synthesising future motions.
The project is currently accepting PhD applications every year.