Computer Science at Aberystwyth University
Research in facial perception has benefited greatly from the ability to alter facial images in a controlled manner, in order to create experimental stimuli. For example, the ability to alter the apparent age, gender or perceived attractiveness allows experimentation into the underlying social, evolutionary and neurological bases of a subject's reactions to facial images.
This talk will describe techniques for creating such stimuli by altering 2D and 3D images using information learnt from groups of images. Textural details play an important role, particularly in capturing and altering age and gender related information such as wrinkles and stubble, and algorithms for processing facial textures will be described. The desire to improve automation and to extend the techniques to video data have led to the development of improved facial feature detection and tracking algorithms and methods for altering the facial dynamics.
Finally, some other applications (facial surgery research, orthodontics and forensic pathology) and possible future directions for the research will be discussed.
Dr Tiddeman's research area is the synthesis and analysis of facial images, for applications in psychology and medicine. The current main focuses of his research are face texture prototyping and transformation, analysis and synthesis of dynamic face image sequences and improved facial age transformation. He was previously at St Andrews University.
