A. Jonathan Howell and Hilary Buxton
This paper is concerned with the types of invariance exhibited by Radial Basis Function (RBF) neural networks when used for human face classification, and the generalisation abilities arising from this behaviour. Experiments using face images in ranges from face-on to profile are presented to show the RBF network's invariance to 2-D shift, scale and y-axis rotation. Finally, the suitability of RBF techniques for future, more automated face classification purposes is discussed.
Download compressed postscript file