This paper prescribes an approach for providing a method of fingerprint authentication from pre-processed fingerprint image
data which are pre-processed by Gabor filter or Haar wavelet transform. The pre-processed data are classified by SVM learning
examples of fingerprint images. The two types characteristic features derived from the pre-processed data have been evaluated.
The first type is minutia matching approach using relation between two minutia and the other is Histograms of Oriented
Gradients approach.