2 open PhD positions in capturing and modelling virtual people from 4D scans, images and video. The PhD position is funded for a duration of 3-4 years. While the computer vision community has seen significant progress in detecting and tracking people in images using deep learning, most approaches are still 2D. The goal of the project is to build the most realistic generative model of 3D people in clothing from real measurements. One part of the project involves “capture”: that means tracking and estimating body shape, soft-tissue and cloth dynamics from data. The other part of the project involves “modelling”: using machine learning techniques the aim is to build models of humans in clothing that generalise to novel human shapes, movements and clothing.
The research will have an impact in several fields such as medicine and artificial intelligence
which requires to track and estimate the human movement and cloth from incomplete sensory data.
The PhD student will work on state-of-the-art research at the intersection between computer vision, computer graphics and machine learning.