5:00pm – 5:45pm: *Making Faces *

Dr Chris Solomon [ University of Kent ]

Facial composites are computer generated images produced by trained police officers from the recollections of a victim or witness to a crime. In general, these systems have a number of shortcomings which mainly originate from the inability of a witness to recall and describe individual facial features. This talk will address the development of a new system which overcomes many of these problems. The system, EFIT-V combines a statistical appearance model of human facial appearance with an interactive evolutionary algorithm. The scientific basis of the system will be described and the system demonstrated. Results from operational use by U.K. police officers will also be presented.

5:45pm – 6:15pm: Break for Refreshments

6:15pm – 7:00pm: *Performance Driven Facial Animations*

Professor Peter McOwan [Queen Mary, University of London]

Markerless motion capture combined with data reduction techniques such as principle components analysis provide a natural means of characterising facial motion and also allow the construction of realistic image-based performance driven animations. This talk reports on work undertaken in collaboration with Alan Johnston in Psychology UCL examining techniques and applications for such animations.