Palmprint Feature Extraction Based On NSCT And LBP

In daily economic life, people's demand of reliable status identification is increasing. Among other authentication methods, biometric recognition has attracted more and more attention as it is the most secure way. As a new biometric recognition technology, palmprint recognition in recent years has drawn a wide public concern and people in the field have begun to devote more time to study it.Palmprint identification can identify individual's status based on the valid information in his or her palm. Compared to other biometric recognition methods, palmprint identification has many advantages including high rate of identification, low-price capture devices, as well as good acceptance of customers. Research shows that the possibility of false rejection for palmprint identification is only one out of 10 thousand cases; and the possibility of false acceptance is one out of 1250 thousand. while the false acceptance rate for fingerprints is one out of 100 thousand. Therefore, the palmprint identification system is competent to meet the harsh and high security requirements.This paper comprehensively discussed palmprint identification technology as well as palmprint identification system. The focus of research is also the key of palmprint identification technology:palmprint feature extraction. This paper also laid a foundation for future research and determined the direction for relevant study. Meanwhile, this paper has limitations, but it raised goals and expectation for subsequent study.Palmprint identification system includes the collection, preprocessing, feature extraction and matching of palmprint. This paper conducts an intensive study of the essential technologies of palmprint identification, and has completed the following works:First, according to analyze the existing feature extraction algorithms, it is found that there are palmprint shifts and rotations during the process of acquiring palmprint image. However, most of the algorithms did not solve this problem very well. As a solution to the above problem, this paper proposed a new algorithm of palmprint feature extraction based on NSCT and LBP. Basic concept, detailed steps and experimental result have been stated in this paper. According to the experiment, recognition rate of this algorithm has reached 99.3%in BJTU_PalmprintDataBase. Compared to other traditional palmprint identification methods, it has higher recognition rate.Second, in order to better apply the proposed algorithm to practical situations, the BJTU Plamprint Identification System has been established. This system adopted the hybrid programming technology of VC++and Matlab. The core algorithm of this system is the algorithm raised in this paper, and has realized the comparison between several traditional palmprint identification methods. This system is outstanding in its stability and expansibility, and it can add new palmprint algorithm without affecting the existing system.