Visual Information Processing research covers a range of topics including vision,
graphics, intelligent behaviour understanding, and biomedical image computing. The work of
the section has led to more than 8 best paper awards at major international conferences
(IEEE FG, ICRA, ISMAR, MICCAI, SensorComm) and attracted four Marie Curie fellows. The
group has pursued a successful strategy of growth in several key areas: In vision, Davison
was awarded an ERC starting grant. Significant achievements include novel modelling and
filtering approaches for SLAM and real-time dense scene mapping. In intelligent behaviour
understanding, Pantic was also awarded an ERC starting grant. This helped to attract an
Imperial JRF (Zafeiriou, now lecturer) and led to: novel approaches to facial action and
emotion prediction as well as novel approaches to robust face alignment, tracking and
expression recognition. The group’s highly successful research in biomedical imaging &
sensing is underpinned two EPSRC programme grants (Rueckert, Yang) and has led to a recent
ERC synergy grant for Rueckert. This success has allowed a significant expansion in the
area of biomedical image computing (Glocker and Lee). Finally, the area of graphics has
been revitalized with the recruitment of Ghosh whose novel methodologies for modelling the
reflectance of materials have had significant impact; for example they have been used for
creating digital faces in the movie Avatar (won Best Visual Effects Academy Award).