This PhD project aims to advance the understanding and simulation of human behaviour by developing innovative methods that integrate visual perception with physical dynamics. The research is divided into two primary areas: understanding human motion and human-scene interaction.
In the first area, the focus is on developing neural networks that accurately and plausibly model human motion. The second area of research aims to create methods for interpreting human-scene interactions. The expected outcomes include sophisticated models that accurately interpret and predict human motions and interactions, contributing to advancements in human-computer interaction, robotics, virtual reality, augmented reality, and animation.