Professor Adrian Hilton
In January 2012 I became Director of the Centre for Vision, Speech and Signal Processing (CVSSP) at the University of Surrey. CVSSP is one of the largest UK research groups in Audio-Visual Machine Perception with 125 researchers and a grant portfolio in excess of £20M. CVSSP research spans audio and video processing, computer vision, machine learning, spatial audio, 3D/4D video, medical image analysis and multimedia communication systems with strong industry collaboration. The centre has an outstanding track-record of pioneering research leading to successful technology transfer with UK industry.
The goal of my research is to develop seeing machines with the visual sense to understand and model dynamic real-world scenes. For example, measuring from video the biomechanics of an Olympic athlete performing a world-record high-jump.
My research is pioneering the next generation of 4D computer vision, capable of sensing both 3D shape and motion, to enable seeing machines that can understand and model dynamic scenes. Recent research introduced 4D vision for analysis in sports, for instance, measurement of football players from live TV cameras. This technology has been used by the BBC in sports commentary to visualise the action from novel directions, such as the referee or goalkeepers view.
I have successfully commercialised technologies for 3D and 4D shape capture exploited in entertainment, manufacture & health, receiving two EU IST Innovation Prizes, a Manufacturing Industry Achievement Award, and a Royal Society RS Industry Fellowship with Framestore on Digital Doubles for Film (2008-11). I am a Royal Society Wolfson Research Merit Award holder in 4D Vision (2013-18) and is currently PI on the S3A Programme Grant in Future Spatial Audio at Home combining audio and vision expertise and InnovateUK project ALIVE, led by The Foundry, developing tools for 360 video reconstruction and editing.
I lead the Visual Media Research (V-Lab) in CVSSP, which is conducting research in video analysis, computer vision and graphics for next generation communication and entertainment applications.
My research combines the fields of computer vision, machine learning, graphics and animation to investigate new methods for reconstruction, modelling and understanding of the real world from images and video.Applications include: sports analysis (soccer, rugby, athletics), 3D TV and film production, visual effects, character animation for games, digital doubles for film and facial animation for visual communication.
Current research is focused on video-based measurement in sports, multiple camera systems in film and TV production, and 3D video for highly realistic animation of people and faces. Research is conducted in collaboration with UK companies in the creative industries.From 2008-12 I was supported by a Royal Society Industry Fellowship to conduct research with leading visual-effects company Framestore investigating 4D technologies for Digital Doubles in film production.
Please contact me if you are interested in current PhD and post-doctoral research opportunities.
Areas of specialism
Virtual and Augmented Reality;
Audio-Visual Signal Processing;
3D Computer Vision;
University roles and responsibilities
- Director, Centre for Vision, Speech and Signal Processing (CVSSP)
- Professor of Computer Vision
- Head of the Visual Media Research Lab. (V-Lab)
CVSSP unveils next-generation augmented and virtual reality technologies for film and television production
World’s first 5G Digital Gaming initiative launches at University of Surrey during the G3: Futures event on 5 July
Contact the press team
Phone: +44 (0)1483 684380 / 688914 / 684378
Out-of-hours: +44 (0)7773 479911
Senate House, University of Surrey
Guildford, Surrey GU2 7XH
Current research is focus on Audio-Visual Machine Perception to enable machines to understand and interact with dynamic real-world scenes. Machine Perception is the bridge between Artificial Intelligence and Sensing (seeing, hearing and other senses). The goal is to enable machines to This research brings together expertise in audio, vision, machine learning and AI to pioneer future intelligent sensing technologies in collaboration with UK industry.
Machine perception of people, and the everyday environments that we live and work in, is a key enabling technology for future intelligent machines. This will impact and benefit many aspects of our lives, from robot assistants able to work safely alongside people, personalised healthcare at home, improved safety and security, safe autonomous vehicles and automated animal welfare monitoring, through to improved social communication and new forms of immersive audio-visual entertainment. The central research challenge for all of these technologies is to enable machines to understand complex real-world dynamic scenes. Reliable Machine Perception of complex scenes requires the fusion of multiple sensing modalities with complementary information, such as seeing and hearing.
Towards this goal I lead research in the area of 4D Vision to understand and model dynamic scenes such as people. Over the past decade we have introduced world-leading technology for 4D dynamic scene capture and reconstruction from multiple view video. This technology underpins future machine perception and is also enabling creation of video-realistic content for film, TV and Virtual Reality.
Our research has pioneered several new technologies including:
- 4D Performance Capture and Animation: Volumetric capture of actor performance enabling video-realistic replay with real-time control of movement. This technology bridges-the-gap between real and computer generated imagery of people. Current research is enabling the use of this technology in Virtual and Augmented Reality to create cinematic experiences.
- Real-time full-body motion capture outdoors: Our research has introduced methods for production quality full-body motion capture in unconstrained outdoor environments (such as a film set) from video + IMU. This technology is enabling content production for film and TV.
- Free-viewpoint video for sports TV production: Research in collaboration with the BBC has introduced technology to produce virtual camera views and 3D stereo of sports events from the broadcast TV cameras. This has been used in soccer and rugby to produce views such as the referees viewpoint or a virtual view along the offside line.
- Video-rate 4D capture of facial shape and motion: The first video system to simultaneously capture facial shape and appearance, used to produce highly realistic animated face models of real people for games and communication. This technology is now used in medicine for analysis of facial shape, in security for face recognition and in the creative industries to create highly realistic digital doubles of actors.
- AvatarMe: A 3D photo-booth to capture realistic 3D animated models of people for games. This technology was used to produce animated models of over 300,000 members of the general public.
This research has received several awards for outstanding academic publications and technical innovation. Research has enabled a number of award winning commercial systems for industrial design, computer animation, broadcast production, games and visual communication.
Current research is pursuing video-realistic reproduction and behavioural modelling of dynamic scenes such as moving people, sports and other natural phenomena. Ultimately the challenge is to capture representations which produce the appearance and interactive behaviour of real people.
If you are interested in pursuing related PhD or post-doctoral research in the area of vision, graphics or animation research contact me for current vacancies.
Collaborators on this research include companies from the film, broadcast, games. communication and consumer electronics industries:
- The Foundry
- 3D Scanners