Two Palmprint Recognition Algorithm Research Based On SIFT And Gabor Transform

With the rapid development of electronic information,automatic identification technology is essential.Palmprint recognition technology as a new biometric technology in recent years,relative to others,the characteristics of palmprint recognition technology are difficult to disguise and hide,more stable,non-invasive and non-contact.In recent years,for the use of palmprint recognition technology in access control,nuclear power station,banking system and quick payment,the research of palmprint recognition will be a promising and practical research topic.Palmprint recognition is the biometric identification technology that based on the main palmprint characteristics to identify palmprint.The key to palmprint recognition is to extract the effective features of palmprint,and design the corresponding classifier for classification and identificationaccording by the extracted palmprint feature.In this paper,the method of palmprint recognition based on SIFT and Gabor wavelet transform is mainly studied from the palmprint segmentation,the palmprint feature extraction and the palmprint classification.The main work of this paper are as follows.Firstly,The original palmprint image is subjected to sequential statistical filtering and binarization and morphological processing,then,split the ROI by the corners which are detected according to the method of Harris algorithm.According to the experimental results,the method has achieved satisfactory results.Secondly,A kind of palmprint recognition algorithm combining Gabor transform and KPCA is resigned.The four-scale and four-direction transformation of the Gabor is performed on the palmprint image,which could get 16 sets of wavelet features.Using the method of KPCA extract the principal components of the normalized features.Combine the obtained 16 sets of principal component characteristics for the final feature vector of each original sample.At last,using the nearest neighbor classifier to identify.Compared with the single Gabor and KPCA algorithms,the algorithm has higher recognition rate.Thirdly,A palmprint recognition algorithm based on SIFT feature matching is resigned.Using the SIFT feature key points of the palmprint match the existing palmprint image SIFT feature key points palmprinted image in the palmprint database,which is using the match statistical characteristics for identify.In the environment of matlab,the simulation results show that the algorithm has a high correct recognition rate.