作者: E. Gomez , C.M. Travieso , J.C. Briceno , M.A. Ferrer
DOI: 10.1109/CCST.2002.1049223
关键词: Computer science 、 Polar coordinate system 、 Pattern recognition 、 Artificial intelligence 、 Computer vision 、 Grayscale 、 Color image 、 Cartesian coordinate system 、 Centroid 、 Biometrics 、 Facial recognition system 、 Contextual image classification
摘要: Biometrics systems based on lip shape recognition are of great interest, but have received little attention in the scientific literature. This is perhaps due to belief that they discriminative power. However, a careful study shows difference between outlines greater than shapes at different images same person. So, biometric identification by outline possible. In this paper obtained from color face picture: image transformed gray scale using transformation Chang et al. (1994) and binarized with Ridler Calvar threshold. Considering centroid as origin coordinates, each pixel envelope parameterized polar (ordered -/spl pi/ +/spl pi/) Cartesian coordinates heights widths). To asses identity, multilabeled multiparameter hidden Markov model used multilayer neural network applied coordinates. With database 50 people an average classification hit ratio 96.9% equal error (EER) 0.015 obtained.