A Palmprint Recognition System Based On The Corner Detection And Competitive Code

In recent years, with the rapid development of mobile digital equipment, mobile phone is providing more and more services to the life of people as a kind of increasingly powerful mobile digital equipment. These services will also involve a large number of users’ privacy and result in increasingly serious security hidden danger. In order to protect the data stored in mobile phone, biometric identification technology is being widely applied to the mobile digital devices. Biometric identification technology apply the body’s physiological or behavioral characteristics to identification, so as to achieve the aim of information and system security protection. Biometric identification include fingerprint features recognition, face feature recognition, iris recognition and the characteristics of palmprint recognition. Among them, the fingerprint identification is the most familiar. So far, there have been many advanced mobile phone support the fingerprint recognition technology. This paper uses a palmprint recognition system based on camera.Compared with the fingerprint identification, palmprint area is larger and contain more abundant texture information, and do not need special fingerprint sensor is the most important. To the palmprint identification, a camera is enough. Compared with face recognition, palmprint feature is more stable and reliable, and the body’s expression, posture, emotion and other factors; Compared with the iris recognition, the cost of palm print acquisition is lower.This paper proposes a palmprint recognition method based on corner detection and the Competitive Code. This method apply the key point to coarsely match a pair of two palmprint images, then match images with the Competitive Code. In this paper, there are three main job: 1) This paper proposes a new method for extracting effective regional of palmprint, the method is based on the gradient direction of the palm contour, accelerated the pretreatment process; 2) This paper apply a method based on SIFT key point detection to extract the direction features of key points, and then coarsely match a pair of two palmprint images with the direction features and the position of the key points; 3)An improved Competitive Code is used to the last matching of the palmprint. In order to verify the effectiveness of the proposed algorithm, this paper adopted Poly U_Palmprint_Datebase data and the palmprint images database from the iphone 5.The experiment proves that this method has very good effect for palmprint recognition.This method studied the features of the SIFT key points in the palmprint image, and proved that some of these features is enough unique and stable to the palmprint matching. The method offers a palm ROI extraction method with the gradient information, speeding up the pretreatment. The Competitive Code is improved to be used to the last matching, make it has good robustness and matching speed. In this paper, the method solved the problem that the complexity of the palmprint recognition algorithm is higher and has certain significance to improve the real-time performance of palmprint identification.