FACE DETECTION AND FACIAL FEATURE EXTRACTION FOR PERSONAL IDENTIFICATION

Abstract

Object detection is a subject coming under the vast topic of computer vision. Personal identification is continuously becoming very important but at the same time it is very difficult task. There are various ways like Passwords and cards widely used for such purpose but they are not fully secure and also they encounter problem of large memory. To carry cards also becomes nonfeasible. Personal identification is also done by identifying some physical features like fingerprints, iris, voice and face. Personal identification based on face identification is used because of no need of physical contact. Face detection and facial feature extraction plays an important role in video surveillance, human computer interaction, and in face recognition. There are various methods used for facial feature extraction and detection like template matching, appearance based methods. This dissertation is considering the image -based approach and implementing effectively the task of frontal face detection in cluttered images, which is a part of face recognition system. This work includes the problems of face variation such as lighting, facial expression, identity, and position. The other variation from object to non -object is treated by training a neural network. The dissertation also includes the view-based face recognition system using linear sub space technique principle component analysis (PCA).