Project Description
This project aims at modelling human motion and body/facial shape data with deep neural network to synthesize virtual characters for computer graphics, animation and games.
The Project
With the advancement of machine learning and artificial intelligence, computer-generated graphics and animation have become more realistic than ever. In this project, you will research on state-of-the-art deep learning algorithms to model human data. This project has a particular interest in 3D human motion and shape, while its scope also covers 2D human image/video, facial expressions, crowd motion, emotion and behaviour, etc. You will then develop graphical applications that synthesize realistic and controllable virtual humans to validate the model effectiveness. Example applications include real-time virtual character controls, virtual clothing try-on, facial make-up synthesis, crowd behaviour simulations. The detailed project direction will be developed during the initial stage of your PhD study according to your experience and strength.
Related past research from our team includes:
- Spatio-temporal Manifold Learning for Human Motions via Long-horizon Modeling, IEEE Transactions on Visualization and Computer Graphics http://hubertshum.com/pbl_tvcg2021deeplearning.htm
- Unified Deep Metric Representation for Mesh Saliency Detection and Non-rigid Shape Matching, IEEE Transactions on Multimedia http://hubertshum.com/pbl_tmm2020mesh.htm
- Interactive Formation Control in Complex Environments, IEEE Transactions on Visualization and Computer Graphics http://hubertshum.com/pbl_tvcg2014formation.htm
Supervision
As a PhD student, you will be supervised by Dr Hubert Shum (http://hubertshum.com/), who is an Associate Professor in Computer Science at Durham University. He has published over 100 research papers in the fields of computer vision, computer graphics, motion analysis and machine learning. He has led funded research projects awarded by the UK Research Council, the Ministry of Defence and the Royal Society. This has facilitated him to supervised 23 PhDs and 6 Post-doctoral Researchers. Engaging both the academic and the industry, he hosted international conferences such as BMVC and the ACM SIGGRAPH Conference on MIG, as well as served as an Associate Editor of CGF and a Guest Editor of IJCV.
During the PhD study, you will receive comprehensive training and research coaching through regular one-to-one meetings with Dr Shum. Such interactive and tailored support can develop your strength and consolidate your research knowledge. Furthermore, Dr Shum’s research team has a supportive culture with team members from all over the world, which provides assistance and collaboration opportunities to each other. These have facilitated his past PhDs to publish their research in prestige journals (e.g. IEEE TIP, IEEE TVCG and IEEE TMM) and to develop their successful career.
Funding Information
This is a self funded project.
Eligibility Requirements
- A research interest in computer graphics and machine learning
- Knowledge in any modern programming languages
- A relevant undergraduate or master degree with good scores
- Good English https://www.dur.ac.uk/learningandteaching.handbook/1/3/3/1/
Application Process
You are encouraged to send an email with a resume to Dr Hubert Shum [email protected] for initial discussions. More information on http://hubertshum.com/
Formal applications should be done online.
Supplementary Information
The position is based in Durham University, which is ranked the 4th in the UK by the Guardian and top 100 in the world by QS Top Universities. As a member of the elite Russell Group, Durham University focuses on research excellence delivered by world-class academics. It is located at Durham in North East England, which is one of the safest cities in the UK with an affordable living cost. The Department of Computer Science is one of the fastest-growing departments in the University, supported by major investments in staff recruitment and a £40m new academic building. 83% of its research outputs are considered as “world-leading” and “internationally excellent” by the UK Research Excellence Framework.
