Research On Palmprint Recognition Technology Based On Wavelet Transform

With the rapid development of computer and Internet technology all over the world, information security has become more important than before. As an important aspect of information security, the reliable and convenient identity recognition technology has invoked increasing interest. Although using palmprint for personal recognition has drawn interests from researchers, it is still a challenging topic.By studying the denoising algorithms, location and normalization methods, feature extraction and matching methods of palmprint image, this dissertation brings forward a series of methods for palmprint analysis.Firstly, a novel palmprint image denoising algorithm-Butterworth-wavelet denoising algorithm is proposed in this dissertation, and it is compared with lowpass filter denoising algorithm, neighborhood average denoising algorithm and Multithr shrink denoising algorithm by experiment. The experiment results prove the validity of this denoising algorithm. Secondly, the methods of LOG operator and 8-neighborhood boundary-tracking algorithm is used to extract palm contour and its chain-code. By analyzing the regularity of chain-code of palm contour, a new and rapid corner-detection method is proposed. Thirdly, the palmprint image is located and the ROI (Region Of Interest) is extracted according to the contour corner. Simultaneously, a ROI extraction method according to the proportion of palm size is put forward. Fourthly, through analyzing the matching capacity of both local wavelet energy feature and global wavelet energy feature, the feature fusion method is proposed .Finally, the validity of palmprint analysis methods proposed in this dissertation is testified by experiments.