Analysis on Texture Feature Extraction Methods for Face Recognition in New Born

Abstract

Biometric methods such as fingerprint, palm print, iris, face and retina are used to detect the persons who involved in the forgery. The face biometric is the most important and widely used in many applications area such as supermarket, railway station, airport, hospitals and other application areas to monitor and control the forgeries. In the present scenario, face recognition gained its importance due to increase in forgeries. To avoid such crimes, face recognition system is given most importance. Feature extraction is an important step for further analysis in face recognition systems. There are many feature extraction algorithms for face recognition systems in the literature. The challenge is to provide better accuracy in face recognition system. While choosing the feature extraction algorithm, we have to consider the parameters which provides better accuracy and less computational time. The features extracted from an image form the basis for classification and the extracted features are used for training and testing purposes. This paper analyses various feature extraction methods such as Local Binary Pattern (LBP), Principal Component Analysis (PCA) and Gray Level Co-occurrence Matrix (GLCM). The Local Binary Pattern generates LBP descriptors. Eigen face and Eigenvectors are computed by Principal Component Analysis and Gray Level Co-occurrence Matrix generates the second order statistical features. These methods are applied on the new born baby images with different expression. The extracted features are given as an input for the Support Vector Machine for classification. The experimental results have shown that the Principal component analysis method provides an accuracy of 91 % and it provided better recognition rate and less computation time when compared with the other feature extraction methods.