Face reconstruction and authentication based on wavelet transform and random forest: A review

Abstract

Face authentication is an emerging technique used in many crime detection systems. Numbers of techniques are available for face identification but most of them are based on front view. This paper gives a novel technique for face detection taking in consideration both front and side view of face. The system takes test image as input and compares it with images in dataset For reconstruction it uses mirror edge generation technique. Before reconstruction it applies preprocessing and curve extraction on input image. Random forest with subset technique is used for authentication. The system is very useful in crime detection if frontal view of criminal is not available.