21.02.2013 Views

Advances in Fingerprint Technology.pdf

Advances in Fingerprint Technology.pdf

Advances in Fingerprint Technology.pdf

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

The next related set of requirements was the description of m<strong>in</strong>utia<br />

distribution us<strong>in</strong>g the same variables and accommodation of variations <strong>in</strong><br />

m<strong>in</strong>utia density along with ridge pattern elements and s<strong>in</strong>gularities. Stoney<br />

and Thornton effectively described the m<strong>in</strong>utia distribution <strong>in</strong> the distal<br />

portion of thumbpr<strong>in</strong>ts, but they assumed a constant m<strong>in</strong>utia density and<br />

did not extend their study to <strong>in</strong>clude the pattern areas of f<strong>in</strong>gerpr<strong>in</strong>ts.<br />

Requirements to <strong>in</strong>corporate two oppos<strong>in</strong>g orientations of m<strong>in</strong>utiae and<br />

three fundamental m<strong>in</strong>utia types were met <strong>in</strong> the model.<br />

Consideration of variation among pr<strong>in</strong>ts from the same source — one<br />

of the most critical requirements — was not seriously addressed by Stoney<br />

and Thornton. Their tolerances for m<strong>in</strong>utia position were derived from<br />

successive pr<strong>in</strong>t<strong>in</strong>gs under ideal conditions: these tolerances are far too low<br />

to apply <strong>in</strong> actual f<strong>in</strong>gerpr<strong>in</strong>t comparisons. Likewise, connective ambiguities<br />

were not factored <strong>in</strong>to Stoney’s ultimate 1.2% ceil<strong>in</strong>g prediction.<br />

Stoney and Thornton also did not meet their f<strong>in</strong>al requirement: provid<strong>in</strong>g<br />

a means to determ<strong>in</strong>e the number of attempted comparisons <strong>in</strong> order to<br />

assess the effect on the chance of false association. The significance of this<br />

omission is questionable, given the criticisms of Robertson and Vignaux. 59<br />

There are thus two fundamental limitations to the Stoney and Thornton<br />

model. First, the survey, and therefore the model, were of very limited scope,<br />

based exclusively on the distal (non-pattern) regions of pairs of thumbpr<strong>in</strong>ts<br />

from 206 males. Second, no mechanism was provided to accommodate pr<strong>in</strong>ts<br />

made under non-ideal pr<strong>in</strong>t<strong>in</strong>g conditions. Stoney and Thornton acknowledged<br />

these weaknesses and their work rema<strong>in</strong>s the only model to have<br />

attempted to describe m<strong>in</strong>utiae with<strong>in</strong> their ridge structure.<br />

Similar to all the other f<strong>in</strong>gerpr<strong>in</strong>t models proposed to date, the Stoney<br />

and Thornton model rema<strong>in</strong>s untested. What is miss<strong>in</strong>g is the use of the<br />

model to predict the frequency of occurrence of m<strong>in</strong>utiae configurations,<br />

followed by the test<strong>in</strong>g of the accuracy of the prediction.<br />

Champod and Margot Model (1995–1996)<br />

Based on the work <strong>in</strong> Champod’s thesis, 11 Champod and Margot presented<br />

a statistical model that utilized computer-generated frequencies of m<strong>in</strong>utiae<br />

occurrence and m<strong>in</strong>utia densities. 60 Subsequently, this model was modified to<br />

reclassify certa<strong>in</strong> additional, closely spaced m<strong>in</strong>utiae as comb<strong>in</strong>ed m<strong>in</strong>utiae. 61<br />

Description of the Champod and Margot Model<br />

Experimental Design<br />

Champod and Margot designed a computer program to search for specific<br />

m<strong>in</strong>utiae on f<strong>in</strong>gerpr<strong>in</strong>t images. One thousand <strong>in</strong>ked f<strong>in</strong>gerpr<strong>in</strong>ts were

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!