AUTHENTICATION AND SECURITY BIOMETRICS BASED ON FACE RECOGNITION

Abstract

Nowadays security becomes a most important issue regarding a spoof attack. So, multimodal biometrics
technology has attracted substantial interest for its highest user acceptance, high security, high accuracy, low spoof attack
and high recognition performance in biometric recognition system. This multimodal biometrics system introduces
recognition of person from two things i.e. face & palm print. Principal Component Analysis (PCA) algorithm is used for
reduction of dimension & extraction of features in terms of eigenvalues & eigenvectors. Feature level fusion technique
used to fuse the results of face & palm prints and then gives the output as per neural network classifier which gives the
correct information about genuine or imposter identity. Automatic person identification is an important task in computer
vision and related applications. Multimodal biometrics involves more than two modalities. The proposed work is an
implementation of person identification fusing face, palm biometric modalities used PCA based neural network classifier
for feature extraction from the face and palm images and hamming distance for calculating iris templates. These features
fused and used for identification. Better result was obtained if the modalities were combined. Identification was made
using Eigen faces, Eigen ears, Template of iris and their features tested over the self created image database.