Archives

A Novel System for Biometric Recognition based on Modification of Fuzzy Inference and Gradient Decent


Abdulkareem Merhej Radhi
Abstract

The ear is an important "biometric" measure in distinguishing people, as it remains constant and does not change over time. This making it an effective tool for distinguishing people. The search for a new direction in the investment of the ear as a means of discrimination has become an urgent and important task. This paper presents a development of the "previously approach adopted by the researcher to recognize people's via the ear approach such that it provides a mathematical model for construction of a reciprocal twin sample to the nearest neighbors and then assembled as a group”. The development depicted in this paper is based on two-dimensional (2D-DWT) conversion to analyze ear image to extract features and using principle component analysis (PCA) to minimize the effect of dimensions of these features. In classification, an ANFIS technique combining the adaptive inference system (ANFIS) with the neural network via Gradient Decent. The experimental results reflect that the proposed system is less complex and achieves the best estimate and less time of calculations where estimated accuracy is 97%.

Volume 11 | 04-Special Issue

Pages: 792-798