Personal Identification by Fingerprints based on Gabor Filters

This work is released in biometric field and has as goal, development of a full automatic fingerprint identification system using methods which have been tested in LAMOSI laboratory. Promising Results of first experiences using radial basis function neural network and support vector machine pushed us to continue the exploitation of new version of neural networks which is spike neural networks and to develop codification and recognition algorithms which are specifically associated to this system. In this context, works were consecrated on algorithm developing of the original image processing, minutiae and singular points localization; Gabor filters coding and testing these algorithms on well known databases such as: FVC2004 databases & FingerCell database. Performance Evaluating has proved that spike neural network achieved a good recognition rate closer to rates achieved by other methods but in a very short time and this make it more useful in online applications.