|
« Back
Indexing Biometric Databases
Biometrics such as fingerprints, face and voice verification are gaining industrial, government and citizen acceptance. The US-VISIT program uses biometric systems to enforce homeland and border security. Governments around the world are adopting biometric authentication to implement National ID and voter registration schemes. FBI maintains national criminal and civilian biometric databases for law enforcement. In spite of the rapid proliferation of large-scale databases, the research community has thus far focused only on accuracy with small databases while neglecting the scalability and speed issues important to large database applications. There is no corresponding research on how the searching and matching techniques scale in terms of speed and accuracy. Even the largest relevant study, conducted by NIST deals with 620,000 fingerprint and 120,000 face images, where as each of the example applications mentioned above require databases which may contain tens of millions of biometric records. The need for the scientific community to address these research problems in handling large biometric databases and the development of reliable methods of securing biometric templates is urgent.
Unlike structured information such as text or numeric data
that can be sorted, biometric data does not have any natural
sorting order. Therefore, indexing presents a challenging
problem. We are working on two different methods for guiding
the search in a biometric database. (i) Binning and indexing
methods that perform a coarse level classification of the
template before performing exhaustive matching. (ii) Clustering
methods whereby natural groups and classes are derived from
the statistical distribution of biometric data. In our preliminary
experiments, we have been able to reduce the search space
to just 5% of the entire database using hand geometry and
signature features.
Publications:
- A. Mhatre , S. Palla , S. Chikkerur and V. Govindaraju V, " Efficient Search and Retrieval in Biometric Databases", SPIE Defense and Security Symposium, Vol - 5779, pages:265-273, March-2005
[pdf..]
- A. Mhatre , S. Chikkerur and V. Govindaraju, "Indexing Biometric Databases using Pyramid Technique", Audio and Video-based Biometric Person Authentication (AVBPA), July-2005
[pdf..]
|
|
 Human iris patterns have a high degree of randomness and uniqueness.
|
