Abstract: In recent years for recognition of person no. of methods are available, out of which Biometric system becomes most reliable one .Unimodal biometric system is one type of biometric system which is widely used in different industry and organizations for authentication of purpose. Even though unimodal biometric system widely accepted for recognition purpose, it has its own limitations. Because of this it is mostly used in those organizations where middle level of security is essential. To overcome limitations of biometric system which is used single biometric traits for identification purpose, there is need of such better performed system. This can be achieved by combining results of multiple unimodal system .So, because of this combining effect, multimodal system has all characteristics of that unimodal systems from which it is build. By considering this advantage of multimodal system over unimodal system objective of proposed system is to build multimodal biometric system which has characteristics as high level accuracy, reliable results of person recognition and robust in nature. It uses finger knuckles, palm print and finger geometry as inputs (biometric traits) for authentication of person. Image of hand acquired through Image acquisition system is used in preprocessing stage to get ROI of fingers and palm. In extraction stage finger geometry and finger knuckles of ring, middle and index fingers are extracted. And also key points which represent palm print features are also extracted to get ROI of palm. KNN classifier is used for classification of purpose. To boost speed of matching process a hierarchical method is used. To get high accuracy of features AND rule fusion is used. With help of thermal and PIR sensors aliveness of person is detected to avoid spoofing in biometric system.
Keywords: Biometric system, unimodal, multimodal, palm print, finger geometry finger knuckles, KNN classifier, AND rule, ROI, Aliveness detection, PIR sensor, Thermal sensor