Research On The Encryption And Decryption Algorithms Based On Non-Contact Palmprint Feature

In the information age, information security has become increasingly important. The foundation of information security is cryptography. The traditional cryptographies are based on password or key which are vulnerable to dictionary attacks, difficult to remember and easy to be illegal stolen. It's just because the lack of relation between the password or key and the genuine users, the system can not distinguish the users'identity. In order to overcome all those problems, we must establish a relationship between the encrypted information or keys and users. However, cryptographic system based on biometrics can overcome the shortages of traditional cryptographic system. The cryptographic technology based on palmprint is an important area of research.At present the most devices used in palmprint cryptographic system are contact. When users use these devices, they have invasions of privacy. And it's easy to spread bacteria and disease, because many people have touched the devices. In order to improve the user's acceptance, a non-contact palmprint capture device is used. The research of the encryption and decryption algorithm based on non-contact palmprint feature has important significance and broad application prospects. This paper's researching contents mainly include the following aspects:Firstly, non-contact devices bring some degree of freedom, when capture the palmprint images. Cause the key points localization is difficult and inaccuracy. For the precision of cryptographic system influenced by the accuracy of key points localization, a localization method of key points based on non-contact palmprint images is presented in this paper, which performs on both finger-separated and finger-closed palmprint images. Finally do experiments by using the palmprint recognition technology. Experimental results show that the proposed method provides considerable performance in both localization accuracy and stability.Secondly, a palmprint cryptography algorithm based on look-up table is presented. The algorithm directly uses palmprint template for data encryption and decryption, and does not require the palmprint features which are used in encryption phase and decryption phase are the same. Only requires that the features are sufficiently similar to correctly complete the data encryption and decryption. The look-up table method is used in encryption and decryption phase, and Reed-Solomon is used to recovery the data. This algorithm effectively eliminates the fuzziness of palmprint template and has high safety.Thirdly, a key generation algorithm based on equal probability quantization method of palmprint feature is proposed. This algorithm can generate exactly the same keys used in encryption and decryption from two similar palmprint features and has user specific. Firstly, determine the quantization parameters, and then use them to quantize any biometric feature vector. The quantization probability is based on minimum error ratio. It also means finding the optimal quantization interval to minimize the sum of FRR and FAR. Then compute the probability in the optimal quantization interval from the global probability density function as the quantization step size. And use the step to equally divide the global distribution domain. The experiments results which are very encouraging demonstrate the effectiveness and commonality of our proposed algorithm.