Scale Invariant Feature Transform Based Face Recognition from a Single Sample per Person

The technological growth has a serious impact on security which has its own significance. The core objective of this project is to extract the facial features using the local appearance based method for the accurate face identification with single sample per class .The face biometric based person identification plays a major role in wide range of applications such as Airport security, Driver's license, Passport, Voting System, Surveillance. This project presents face recognition based on granular computing and robust feature extraction using Scale Invariant Feature Transform (SIFT) approach. The Median filter is used to extract the hybrid features and the pyramids are generated after the face granulation. Then, DoG pyramid will be formed from successive iterations of Gaussian images. By this granulation, facial features are segregated at different resolutions to provide edge information, noise, smoothness and blurriness present in a face image. In feature extraction stage, SIFT descriptor utilized to assign the intersecting points which are invariant to natural distortions. This feature is useful to distinguish the maximum number of samples accurately and it is matched with already stored original face samples for identification. The simulated results will be shown used granulation and feature descriptors has better discriminatory power and recognition accuracy in the process of recognizing different facial appearance.