RUS  ENG
Full version
JOURNALS // Upravlenie Bol'shimi Sistemami // Archive

UBS, 2012 Issue 38, Pages 65–77 (Mi ubs607)

Information Technology Applications in Control

Statistical recognition of faces based on feature points geometry for transport security systems

A. N. Karkishchenko, I. Grechuhin

JSC Research and Design Institute for Information Technology, Signalling and Telecommunications on Railway Transport

Abstract: We consider a formal statement of the face feature points construction problem in terms of statistical identification. The process of feature points construction based on the training set is described. The formal definition of statistical classes is given. We suggest a method for face identification using statistical classes and feature points. In conclusion the results of experiments and comparison with the known methods are given.

Keywords: biometric identification, principal component analysis, Mahalanobis distance, active shape model, statistical class, measure of inclusion.

UDC: 004.93'1
BBK: 05.13



© Steklov Math. Inst. of RAS, 2024