A Multi Model Based Bio Metric Analysis Using Gaussian Distribution

Abstract

Most of the  bio metric analysis in the industry till now is using unimodel verification. major Systems has to deal with added demands such as coverage of vast data base and demographic variety, varied use  in different environment, and demanding quality requirements. As the unimodel with limited features cannot fulfill all the need it better to add the other features also which are available at those places. A multimodal  biometric system combines features from multiple biometric traits, sensors, algorithms, and other components to make a identification conclusion along with enhancement of accuracy, the fusion of biometrics can also come over the problems like spoofing, limited population etc..there is huge research work is going on the these multimodal but there is huge gap in solving the problems like incompleteness of the features to compare.  Therefore, here we developed a hybrid multimodal biometric authentication approach fusing palm print and fingerprint traits at score-level. here in the multi model authentication,  Generalized Gaussian distribution is used on the features of both the prints which are further used for the verification and validation. here we have taken criminal mapping application with huge data base. The model performance is measured using the FAR and MDR