Study Of Palmprint Feature Extraction Algorithm Based On Multi-resolution Analysis And Gray Level Co-Occurence Matrix

In recent years, with the rapid development of information science and technology,information security is facing unprecedented challenges. Biometric recognition uses thephysiology and behavior characteristics of the human body for authentication, providing areliable guarantee for the solution of information security. At present, biometric recognitiontechnology has been developed maturely and used in many fields widely.Because the palm has the characteristics of large area, more details and having enoughidentifying information though the palm is damaged or missing, palmprint recognitiontechnology has become the key research topics of many universities and research institutionsin recent years. In addition, compared with fingerprint identification, the current palm-basedbiometric technology research and application is far less wide .Therefore, whether from theview of scientific or from the perspective of market needs, we can all know it has theimportant theoretical value and social value of doing research about palmprint recognition.The core of this paper is the palmprint feature extraction. The main work is as follows:(1) In the first two chapters, the paper introduces the basic situation of biometricidentification technology and palmprint recognition separately, and focuses on the details ofthe preprocessing and feature extraction of the palmprint recognition process.(2) To solve the problem of the palmprint image's position and angle offset, and improvethe palmprint image's contrast, the paper pretreat the palmprint image, including binarizing,positioning, normalizing and Image enhancing. It provides for the next palmprint featureextraction.(3) After pretreatment, the palmprint image is decomposited using wavelet. Then thepaper calculates the gray level co-occurrence matrix for all levels of sub-image,and gains theeigenvectors which can reflect the features of the palmprint image comprehensively. So it achieves the study of palmprint feature extraction algorithm based on Multi-resolutionAnalysis and Gray Level Co-Occurence Matrix.(4) By the method of artificial matching , we calculated the identification accuracy of thepaper's algorithm. Then we compared it with the identification accuracy of the algorithmbased on Gray Level Co-Occurence Matrix. So the effectiveness and superiority of thispaper's algorithm is verified.