Unregistered Face Discrimination by the Face Orientation and Size Recognition
Kiyomi Nakamura, Hironobu Takano · IEEE International Conference on Neural Networks/IEEE ... International Conference on Neural Networks · 2007
Emulating the parietal cortex, a "rotation and size spreading associative neural network" (RS-SAN net) was developed. Using the RS-SAN net, a new personal authentication method was proposed which was not influenced by the coplanar rotation and size changes of the input faces. The recognition characteristics of the RS-SAN net for both learned (familiar) and un-learned (unfamiliar) face images were investigated. The RS-SAN net had fairly good orientation and size recognition characteristics only for learned faces, but not for unlearned faces. The recognized orientation and size for unlearned faces were heavily dispersed from those of input face although the orientation and size for learned faces were concentrated around learned one. By using the unique characteristics of orientation and size recognition, new recognition method with the imposter detection using recognized orientation and size was developed. The experimental result obtained by new recognition method indicated the false acceptance rate drastically decreased. The imposter rejection method using recognized orientation and size provided the effective improvement of face recognition performance.