Research And Implementation Of Palmprint Identification Algorithm

Palmprint identification technology uses the texture information of human's hand as the biological characteristics to automatically identity, which is the new technology in the area of biometric identification. Currently palmprint identification is used in banks, hospitals, libraries and police departments, such as the smart card authentication system, which has good market prospects and industry prospects.This paper at first uses digital camera to collect the right palmprint images, and then preprocesses the collected palmprint images and extract the feature of the palmprint images. In the first stage through extracting the feature, it establishes the image database; in the second stage, using the extracted feature of the palmprint images through wavelet energy method to compare with the database and match in the database to form a complete offline palmprint identification system. The identification system is through the Visual C++to realize on the PC.In the design process we have done the follow work points:(1) on the basis of the original binarization algorithm, increasing the filter of binarization makes the palmprint image after binarization more clearly. (2)bringing forward a new edge detection algorithm, which haves a good effect on the palmprint image after binarization and computes in a simple way. (3)perfecting the traditional algorithm extracting the key points from palmprint images, to make the final key points which are extracted from the palmprint images unique. (4)on the basis of the original transforms of angle, increasing the determination algorithm,to make the edge of the rotated images relatively smooth.In the paper, the palmprint identification system is based on VC6.0 development platform to build on. Through the 320 palmprint image-set identification test, the recognition rate can reach 75 percent. And the structure of system is clear, operating interfaces is friendly and the system speed is rapid and so on.