| Feature extraction in
fingerprint images by learning templates
Presented by: Xuejun Tan
ABSTRACT : Most current techniques utilize complex preprocessing
and postprocessing methods in fingerprint feature extraction.
And all these methods heavily depend on many experimental
parameters. In this presentation, I proposed a learning algorithm,
whcih is based on GAs, to characterizes the features in fingerprint
recognition. The performance of the learning algorithm is
evaluted by so called 'Goodness Value' and the performance
of a identification algorithm. Experimental results show that
the proposed algorithm is robust and efficient, it can avoid
complex preprocessing and postprocessing.