Artificial neural network based automatic face parts prediction system from only fingerprints

Abstract

Biometrics is a deeply studied and highly developed technology. While biometric systems have been used primarily in limited applications requiring high security tasks like criminal identification and police work, more recently they have been receiving increasing demand for person recognition applications. In spite of all developments in biometrics, there is no study investigating relationships between biometric features in the literature. This study presents a novel intelligent approach analysing the existence of any relationship among fingerprints and face parts. Proposed approach is based on artificial neural networks. Developed system generates the stationary face parts of a person including eyebrows, eyes and nose from only one fingerprint image of the same person without knowing any information about his or her face with the errors among 1.4 % and 4.8 %. The satisfactory results have indicated that there are close realitionships among fingerprints and faces. Improving of the proposed system is still sustained for the purpose of analysing and modelling of this relationship for the future developments in biometrics and security applications.