Image and Sensory
N. Costen, Dr B. Li,
Dr E. Prakash and Dr D. Britch
110 journal papers and
international conference contributions since RAE 2001. At this and
other institutions, 2 Research Fellows and 7 PhD students have been
The personnel of this group
combine strengths in image and sensory modelling and data analysis for
a range of applications. The group works on face and voice
interpretation, human motion analysis and reconstruction, feature-based
algorithms, 3D modelling and visualization, computer animation, games
technology and novel data analysis. Strong links exist with many
external groups including Manchester,
Aberystwyth, MIT, SUNY-Stony Brook, UIUC and Nanyang
Technological, Beijing, Zhejiang
and Beijing Normal
and also with the Intelligent Systems Group at MMU.
To advance the theory and
practice of computer vision and pattern recognition, 3D modelling and
visualization, data analysis and optimization.
To develop innovative models
and algorithms for rapid 3D human modeling,
interactive cinematic quality rendering and synthesis of natural
To develop robust real-time
algorithms for human face and body tracking, recognition and behavior understanding from video and cognitive
models of these processes.
To combine the strengths and
expertise of group members, extending research and application areas
incorporating facial expressions for behavior
understanding and HCI, low bit-rate multi-modal realistic facial
synthesis and digital characters.
- Establishment of links and
formal collaborations with international groups, underpinned by a solid
expression synthesis received the SIBGRAPHI Conference Best Paper
Award; published in the Computer
Graphics Forum, the leading international computer graphics journal.
representation and analysis of human faces (images 3D range images
and videos); published as an invited paper in the International
Journal of Imaging Systems and Technology.
- Self-initialising tracking
algorithms to reconstruct articulated motion from sparse features;
published in IEEE
Man and Cybernetics.
person specific face model, providing high fidelity expression
animation; published in IEEE
Transactions on Visualization and Graphics.
- Novel motion-based
frequency-domain human periodic movement recognition model; published
and Vision Computing.
of novel sparse classification algorithms using 1-norm penalty
functions, thus solving the feature selection problem; published in
- Research on feature-based
non-rigid articulated motion, reconstructing human motion for animation/game
projects. Reported by a variety
of outlets including the BBC
funding (with Manchester
for £413,000 (£190,000 to MMU) for human and computer face
recognition from video sequences.
- EPSRC and Microsoft funding
of £85,000 to develop novel
software architectures and algorithms for usability in console,
online and mobile games technology.