Hao Hu is a Ph.D. student in the lab of Tim Murphy at UBC. His project focuses on learning the representation of behavioral time sequence at a low dimensional manifold using recent machine learning techniques.
Synth-mouse is a project which aims to automatically estimate the library of all possible animal pose and its frequency without extra human labor. We plan to learn basic knowledge about animal movements by simulating synthetic mice performing similar movements as real mice movement videos. This project will facilitate ethological feature tracking and improving mouse behavior classification algorithms.
Check out Hao’s presentation of his CONP project in a short talk entitled “Synth-Mouse: A library of mouse 3D poses from synthetic-real data exploration”.