A New Intelligent System for Predicting Gender from Fingerprint

Abstract

This paper proposes a new method for predicting genders from only fingerprints based on Artificial NeuralNetwork (ANN) model. The modelling tasks are achieved by capturing fingerprints, analyzing an obtainedfingerprint part (for example cropping 5x5mm part from a fingerprint image), determining ridges crossed by thediagonal of the obtained fingerprint part automatically, finding the ridge counts, ridge thicknesses, averagefingerprint ridge counts of individuals, combining these features in a model, preparing and setting the structureof it and finally testing the model. The results of proposed model have shown that the best model achieves thetask within 72% accuracy.