Analysis and Design of Principal Component Analysis and Hidden Markov Model for Face Recognition

Biometric detection is considered as an important tool for states to use to strengthen the safety measures. Biometric increases robustness of the biometric system against many attacks and solve the problem of non-universality. Since facial image is the mandatory biometric identifier this proposed work focuses on the use of facial image. Face authentication involves extracting characteristics set such as eyes, nose, mouth from a two dimensional image of the user face and matching it with the templates stored in the database. Facial recognition is a difficult task because of the fact that the face is variable social organ which displays a variety of expressions. The proposed method is for facial recognition for both images and moving video using Principal Component Analysis (PCA), includes Hidden Markov Model (HMM) technique and Gaussian mixture model (GMM) and Artificial Neural Network (ANN), Since HMM technique is a powerful tool for statistical natural image processing and videos. PCA is a statistical procedure which uses an orthogonal transformation. Face recognition techniques dependent on parameters like background noise, lighting, eyes moments, lips and position of key features. Moreover, the face patterns are divided into numerous small scale states and recombined to obtain the recognition result. The experimental results are obtained from this proposed work has been achieved the performance parameters 99.83% of false rejection rate (FRR) and 0.62% of false acceptance rate (FAR) and an accuracy of 96% is implemented using Matlab2012A. © 2015 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the International Conference on Nanomaterials and Technologies (CNT 2014). * Corresponding author. Tel.: +91-8722497866, +91-9880994563. E-mail address: [email protected], [email protected]