Palmprint Recognition Technology Research Based On Hash Algorithm

Palmprint recognition is the technology which identifying by palmprints.After years of research,scholars both at home and abroad have made many achievements based on the theory analysis and single palmprint identification technology applications.However,with the enlargement of palmprint database,palmprint recognition system will exist problems like slow.Therefore,the research of fast-identification technology aimed at large palmprint database has great academic and utility value.In recent years,due to the characteristic of hash algorithm(Hashing)which has obvious dimension reduction effect,therefore,this algorithm has been popularized to the research of image fast retrieval from original image encryption application.The subject aims at automation of Chinese academy of sciences(CASIA)large palmprint database,with the research goal of palmprint identification technology based on hash algorithm.The specific research work are as follows:First,aimed at CASIA palmprint database,research on palmprint image preprocessing,including image binarization,morphological analysis,contour extraction and region of interest(ROI)extraction process,such as to get the corresponding ROI image;Second,in view of the ROI image,adopt the method of combining the Competitive Code with hash algorithm.First of all,extract features by the method of Competitive Code based on coding.Then,make use of the dimension reduction thought of hash algorithm.Finally,classify by matching the angular distance.Third,in view of the ROI image,adopt the method of combining PCA with hash algorithm.First of all,extract features by the method of principal component analysis(PCA)based on subspace.Then make use of the dimension reduction thought of hash algorithm.Finally,classify by Euclidean distance.Fourth,palmprint recognition authentication system is developed in VC++ 6.0 platform,experimental verified on the CASIA palmprint database.The results showed that,by comparing with each other,the average recognition rate of the PCA Hashing method is slightly lower than the PCA method.And there is little difference in average recognition rate between the Competitive Code Hashing method and the Competitive Code.Yet the recognition time is both shortened about one fifth.It can be seen,the above two methods employed herein on the basis of the library to ensure large palmprint correct recognition rate,it is preferred to solve the problem of slow speed of recognition.