Small Area Fingerprint Encryption Algorithm Based On Fingerprint Mosaic And Feature Descriptor

With the development of portable electronic devices driven by embedded technology,the application of fingerprint identification technology in mobile intelligent terminal devices has been popularized,and fingerprint encryption in those devices is particularly urgent to meet the application more secure.As the most widely used biometrics,the fingerprint combined with encryption technology overcomes the disadvantages of the traditional information encryption technology,which provides the new ideas and solutions for the key protection and the biometric template security.The fingerprint fuzzy vault algorithm is one of the best known fingerprint encryption algorithms at present.It binds the personal identity information with the key by using fingerprint characteristics,and perfectly reconciles the contradiction between the fuzziness of fingerprint identification and the accuracy of encryption technology.However,compared with the conventional fingerprint scanner,the fingerprint image area acquired by the portable electronic devices is greatly reduced,which directly leads to the shortage of fingerprint characteristic information that can be extracted and used,and makes it difficult to directly apply the fingerprint fuzzy vault algorithm to the fingerprint image with small area.This paper has done some innovative work based on the fingerprint fuzzy vault algorithm for small area fingerprint,and the main contents include the following aspects:(1)This paper designs and implements a small area fingerprint fuzzy vault scheme based on fingerprint mosaic.Aiming at the problem that the fingerprint feature information extracted is less,the fingerprint mosaic technologies on feature-based and image-based are presented to increase the small-size fingerprint information respectively to meet the basic requirements of the fuzzy vault algorithm.(2)Aiming at the fingerprint matching problem on encryption domain,a fuzzy vault algorithm based on the fusion feature descriptor is proposed in this paper.In this algorithm,the fingerprint feature descriptor is reconstructed,and the Helper Data is used as the breakthrough point to construct the fingerprint fusion feature descriptor,which improves the similarity between the detail point pairs and the fingerprint matching accuracy on the fuzzy vault algorithm's encryption domain.In this paper,FVC2002 DB1 fingerprint database and XDFinger small area fingerprint database were compared.Experimental results show that compared with the conventional fingerprint,the fingerprint fuzzy vault is difficult to be applied to small area fingerprint.In order to verify the validity of the proposed methods,more experiments have been carried out on the XDFinger database.The experimental results show that for small area fingerprint images,the failure-to-capture rate(FTCR)is reduced from 38% to 6% and the genuine accept rate(GAR)is increased from 46% to 68%.Compared with the traditional fuzzy vault algorithm,the fuzzy vault algorithm on the fusion feature descriptor can increase the GAR from 60% to 75% and reduce the false accept rate(FAR)from 0.12% to 0.02% on the premise of ensuring the same security.To sum up,the performance of the fingerprint encryption algorithm proposed in this paper has been significantly improved in small area fingerprint,which to some extent improves the security performance of mobile payment technology based on fingerprint identification technology,and promotes the application of small area fingerprint encryption algorithm in smart mobile terminals.