Master Thesis - Department of Computer Science
Master Thesis - Department of Computer Science
Master Thesis - Department of Computer Science
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
combine the discriminative informations obtained from null space and range space<br />
for constructing a dual space. Then forward or backward selection technique is used<br />
to select optimal feature set from dual space. On the other hand, decision fusion<br />
based method constructs two different classifiers on the null space and range space<br />
separately and then combines them using decision fusion strategies. Results <strong>of</strong> all the<br />
three methods have been shown over three standard databases.<br />
There are severe drawbacks and risks <strong>of</strong> using a unifeature biometric recognition<br />
system, specifically recognition based on fingerprint or face alone. A machine which<br />
has the ability to only analyze data, but has no power to determine if the data has<br />
been fed into by an authorized person or not, is vulnerable to such acts <strong>of</strong> impostors<br />
who are always on the lookout to create innovative means <strong>of</strong> breaking into a system.<br />
Multimodal biometry which uses multiple sources <strong>of</strong> information for decision-making<br />
is a first choice solution for the above defined problem.<br />
In this thesis, a new approach for combining evidences from face and fingerprint<br />
classifiers at the decision level has been proposed. In this approach, each <strong>of</strong> the face<br />
and fingerprint classifier is separately exploited on the basis <strong>of</strong> availability <strong>of</strong> class-<br />
specific information to enhance combination performance. Results using face and<br />
fingerprint databases, show that the proposed methodology <strong>of</strong> using class-specific<br />
information at classifier’s response outperforms the state-<strong>of</strong>-art fusion techniques.<br />
Keywords: Face, dual space, subband face, decision fusion, feature fusion,<br />
eigenspace, multimodal, biometry.<br />
xix