Privacy Preserving Biometric Verification

Template and databases are very important part of biometric system and attacker mostly attacks on template and database of biometic system, so securing them is very important. Size of the database is very large so reduce database size is also very crucial issue these days. Our aim is to provide privacy to the user by protecting template and to reduce dimension of the template database for effective computations. With this aim in mind we have proposed Classical Multidimensional scaling technique which is very effective in dimension reduction as well as it also provides privacy to the user by securing template. We have implemented biometric system using PCA, RP, PCARP and Classical MDS. Experimental results show that proposed technique is more effective in dimesion reduction as well as it also improves the recognition rate of the biometric system. Comparative analysis of Proposed method with PCA, RP, and PCARP are also discussed.