Special

Multimodal Biometric System at IIT Kanpur

The project sought to design and develop a multimodal biometric system, which makes use of multiple biometric traits under the open source platform and fuses them in order to meet stringent performance requirements. This enhanced structure takes advantage of the proficiency of each individual biometric (face, finger, iris, etc.) and can be used to overcome some of the limitations of a single biometric. The system has considered five traits, i.e., face, fingerprint, ear, iris and signature, and can be easily scalable and portable on any open platform.

PROCESS->

Most biometric systems deployed in real world applications are unimodal, that is, they rely on the evidence of a single source of biometric information for authentication (e.g., a fingerprint or face). Some of the limitations imposed by unimodal biometric systems can be overcome by including multiple sources of biometric information of establishing identity.

Such systems, known as multimodal biometrics systems, are expected to be more reliable due to the presence of multiple, fairly independent pieces of evidence. They address the problem of non-universality, since the use of multiple traits ensures that more users can be accommodated in the system. Multimodal systems also deter spoofing since it would be difficult for an imposter to spoof multiple biometric traits for a genuine user simultaneously.

The system has been tested on database prepared at IIT, Kanpur, and is giving an overall accuracy of 99.23 per cent.

IMPACT ->

Currently, this is the first open source software for multimodal biometrics. The system relies on multiple traits and works even if few trait inputs are not present. It is based on multimodal traits and new traits can easily plugged-in. So it is easily scalable.

SCALABILITY ->

The multimodal system developed is scalable and new traits can be easily plugged-in to it. Since the system has been developed on Java and MySQL, it can easily be implemented on any open system and can work on large database. The developed system can easily be customised based on the users' requirements, the degree of security, availability of data and resources. There are plans to add new traits to the system to improve its reliability and robustness. Individual recognition algorithms used in various traits are being enhanced to improve their robustness.