The Research On A Palmprint Recognition Method Of Gabor Wavelet And K-L Transform

Biometrics is the technology that uses human body characteristics for automated personal authentication. Personal characteristics including physiological characteristics (such as fingerprint, iris, face, palmprint) and behavioral characteristics (such as pace, signature) are universal, unique and permanent, they are also unforgettable and hard to imitate and forge, so that they can be used to reinforce or even replace the traditional authentication tools. Biometric technology will be widely used in national security system, financial system, e-commerce and other related field.Palmprint recognition technology makes use of human palm texture as biometric features to identify human identity. Comparing with other biometric features such as fingerprint, face and iris, palmprint recognition has become a very attractive and best development potential biometric technology by its'abundant information, stable features and easy collection. But, nowadays, the research of palmprint recognition at home and abroad is not as extensive and deep as that of fingerprint recognition, face recognition and speaker identification. According to the detail characteristics of palmprint, this paper makes a deep research to the key technology and core algorithms in palmprint recognition system.The current developmental situation on palmprint recognition is briefly introduced in the paper, and then each element in palmprint recognition system and its'key technology is discussed in detail, in this part, Arithmetics mainly used in palmprint recognition in the recent years are studied detailly. In this way, four kinds of arithmetics used in palmprint feature extraction in exitent literatures are summarized. The paper puts emphasis on analysising the following feature extraction methods, Gabor wavelet transform, K-L transform, Fisher linear discriminant, generalized K-L transform, then according to their advantages and disadvantages, banding together the feature extraction method based on time to frequency field transform and the feature extraction method based on subspace, a novel algorithm of palmprint feature extraction using Gabor wavelet and original space discreption information K-L transform is put forward. The arithmetic makes use of Gabor transform to obtain Gabor eigenvector of palmprint from palmprint image that has been processed, and then the original space discreption information K-L transform is used to extract the master components and reduce dimensions of Gabor features. The arithmetic makes full use of Gabor function's fine capability of feature extraction and also effectively solves the problem of feature's high dimensions, by experimenting in the PolyU Palmprint Database offered by Hongkong Polytechnic University and self-captured palmprint database, it could be proved that the proposed method has good robust and effectiveness, and could well realized palmprint feature extraction.