Please use this identifier to cite or link to this item:
|Comparative Study of Normalization Techniques Used for the Fusion of Different Biometric Traits
|Singla, Sunil Kumar
|Unimodal Biometric;Normalized;False Rejection Ratio
|These days biometric are one of the fundamental technologies that are being used for individual authentication in many industrial, confidential and domestic applications. When we consolidate the information from one biometric trait, is known as unimodal biometric. However unimodal systems are reliable, still the popularity of these systems is degrading due to the problems such as noisy sensor data, non-universality and spoof attacks. These limitations can be overcome if we fuse more than one trait to make a system called Multibiometric system. The research work demonstrated in this thesis describes fusion of the multiple biometric traits at score-level. The research work has been stressed around only identification mode i.e. fusion of face, left index finger and right index finger biometrics in the two recognition modes of verification and identification. The experimental approach in this study is also extended to the combination of transformed (normalized) score with fuzzy logic fusion. The results obtained after the experiment depict that, integrating the information at score level is more effectual than the use of single biometric trait. The thesis confronts an exhaustive description of the systematic investigation to establish facts undertaken for the purpose of authentication. This exhibits the consequences on the basis of experiment in terms of False Rejection Ratio and Genuine Acceptance Rate and provides a broad investigation on a moderate database. In this thesis, the problem of score level fusion related to different range of data has been addressed. Due to the heterogeneous nature matching scores of the various biometric traits, score normalization is required to transform these scores into a common domain prior to fusion. The normalization schemes have been evaluated both in terms of their efficiency and robustness to the presence of outliers in the training data. We have developed an experimental environment based on transformation for the successful execution of the plan. Further, a fuzzy logic based fusion system has also been developed to fuse three biometric traits. The advantage of using fuzzy logic is its ease of use and flexibility. With the help of fuzzy logic based system 91.3% genuine acceptance rate has been achieved.
|Master of Engineering (Electronic Instrumentation and Control)
|Appears in Collections:
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.