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 issues of sensor interoperability and quality in the context of fingerprints and makes a three-fold contribution. The first contribution is a method to facilitate fingerprint sensor interoperability that involves the comparison of fingerprint images originating from multiple sensors. The proposed technique models the relationship between images acquired by two different sensors using a Thin Plate Spline (TPS) function. Such a calibration model is observed to enhance the inter-sensor matching performance on the MSU dataset containing images from optical and capacitive sensors. Experiments indicate that the proposed calibration scheme improves the inter-sensor Genuine Accept Rate (GAR) by 35% to 40% at a False Accept Rate (FAR) of 0.01%. The second contribution is a technique to incorporate the local image quality information in the fingerprint matching process. Experiments on the FVC 2002 and 2004 databases suggest the potential of this scheme to improve the matching performance of a generic fingerprint recognition system. The final contribution of this thesis is a method for classifying fingerprint images into 3 categories: good, dry and smudged. Such a categorization would assist in invoking different image processing or matching schemes based on the nature of the input fingerprint image. A classification rate of 97.45% is obtained on a subset of the FVC 2004 DB1 database.