Certain investigations on Feature extraction classification And recognition for shoeprint images

Abstract

Recognition or identification of images is the current focus in the newlinefield of security especially in forensic science application There is a need to design a system that is helpful in recognition of shoeprint images with fast speed and less in time For the past two decades many techniques for the analysis of images have proposed in the literature for the purposes of feature extraction recognition and classification Such approaches include the properties of individual pixel elements using statistical features Earlier finger prints iris palm prints face recognition and DNA profiling plays a major role in security and forensic science Nowadays shoeprint recognition becomes more popular in pattern recognition The system consists of two parts The first one is preprocessing including orientation estimation normalization locating the shoeprint region and database generation the second phase starts with feature extraction and ends up with the classification scheme Feature extraction is a challenging problem where research is done for a long time This thesis explores the use of various wavelet transforms and Principal Component Analysis PCA to the task of feature extraction and proposes improvements in this area newline newline