Given the statistic showing that approximately 75% of the faces in home photos are non-frontal, capability of dealing with multi-view faces becomes more and more important for many face-related applications. Multi-view face detection thus becomes a challenging problem due to due to large amount of variation and complexity brought about by changes in facial appearance, lighting and expression. Changing in facial view/pose complicates the situation because the distribution of multi-view faces in a feature space is more dispersed and more complicated than that of frontal faces. In this project, you will be study the state-of-the-art technologies in multi-view face detection and develop a small system to demonstrate the efficiency of using multi-view appearance model in face detection. You will be working at this project closely with a small group in NICTA.
