Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Lane Department of Computer Science and Electrical Engineering

Committee Chair

Arun Ross.


This thesis addresses the problem of biometric indexing in the context of fingerprint and iris databases. In large scale authentication system, the goal is to determine the identity of a subject from a large set of identities. Indexing is a technique to reduce the number of candidate identities to be considered by the identification algorithm. The fingerprint indexing technique (for closed set identification) proposed in this thesis is based on a combination of minutiae and ridge features. Experiments conducted on the FVC2002 and FVC2004 databases indicate that the inclusion of ridge features aids in enhancing indexing performance. The thesis also proposes three techniques for iris indexing (for closed set identification). The first technique is based on iriscodes. The second technique utilizes local binary patterns in the iris texture. The third technique analyzes the iris texture based on a pixel-level difference histogram. The ability to perform indexing at the texture level avoids the computational complexity involved in encoding and is, therefore, more attractive for iris indexing. Experiments on the CASIA 3.0 database suggest the potential of these schemes to index large-scale iris databases.