Face recognition with region division and spin images

This paper explores how spin images can be constructed using shape-from-shading information and used for the purposes of face recognition. We commence by extracting needle-maps from gray-scale images of faces, using a mean needle-map to enforce the correct pattern of facial convexity and concavity. Spin images [4] are estimated from the needle maps using local spherical geometry to approximate the facial surface. Our representation is based on the spin image histograms for an arrangement of image patches. We demonstrate how this representation can be used to perform face recognition across different subjects and illumination conditions. Experiments show the method to be reliable and accurate, and the recognition precision reaches 98% on CMU PIE sub-database.