Using High Performance Computers to Match Fingerprints

Computers can be very good at matching. Ebay matches millions of buyers and sellers. Millions of people find romance at eHarmony.com. Yi Chen, a PhD candidate in Computer Science and Engineering, studies a much more complex problem: matching fingerprints. She uses the High Performance Computer Center at Michigan State University to do this research.

Chen’s work is one of a number of projects underway at MSU’s Biometrics Research Lab. (http://biometrics.cse.msu.edu) . She is working on a thesis entitled “Automatic Fingerprint Recognition Using Extended Feature Set.” She also serves as a research assistant in the biometrics lab.

“In biometrics you want a large database” Chen explains. In one study she analyzed 2000 fingerprint images each at 1000 dots per inch resolution. She describes this as a “huge dataset” because the number of possible matches one has to perform is of the order of 1 million. She needs the large memory capacity and powerful processing resources in order to compare so many images at such high resolution.

Her software seeks to match fingerprints by analyzing and comparing image features such as ridge contours and pores. She works with a new standard from the National Institute of Standards and Technology (NIST) that allows researchers in the field to exchange fingerprint data with an extended feature set. The 1000 dot per inch resolution is necessary to analyze at the pore level, looking at features as small as 60 microns.

Traditionally fingerprint comparisons involve two dimensional images. Part of Chen’s research involves the pioneering field of 3D fingerprint matching. This literally new dimension of course increases the size of the already-large datasets even further.

Chen has given presentations on her work at major biometrics conferences in the United States and in Europe.

In 2007 Chen won one of 10 graduate research fellowships offered annually by the National Institute of Justice.

Chen’s research is under the supervision of Dr. Anil K. Jain, University Distinguished Professor and head of the Biometrics Lab. Dr. Jain and collaborators have published numerous journal articles and he is co-author of a recent book on biometrics.

Asked when she first found computers exciting, she said it was the first time she found her dad’s name in Google.