Exploration of fingerprints for finding gender information using machine learning

Gender identification plays an important role in many applications, such as in criminal investigations, to reduce the comparison of the target data, automate the preparation of the content, suggest advertising, automate the male/female only place entrances, and many more. Most of the biometric systems used to identify gender are designed based on fingerprint data because of its reliability, low-cost, and easy enrollment when compared with the iris, voice, and palm print. In this thesis, the artificial neural network approach was used to extract the machine learning model, where the direction map of the ridges in the fingerprint was used as a feature to determine the gender of the fingerprints. In addition, a set of data belonging to each of the ten fingers was formed. Next, the models were trained and tested. Finally, the results were compared with each other. As a result, the pinky finger consisted of more gender information when compared with the remaining four fingers, for both the left and right hand, for the used datasets.