Biometric Driver Identification 2008-01-0196

This paper deals with the use of biometric methods of driver identification in the vehicle. It compares different methods and evaluates their possible uses in the vehicle. The methods of fingerprint, iris recognition and face recognition are dealt with in particular.

The second part of the paper describes the implementation of a face recognition system in the vehicle. Important system components are a camera, preferably mounted in the instrument cluster or on the steering column, IR lighting and a computer unit that implements the necessary algorithm. The system observes the driver as he or she enters the vehicle, chooses the images suitable for identification from the image sequence and is able to re-recognize a person it has “seen” before.

This paper focuses on integration in the vehicle with regard to the vehicle-specific environment and connection to the vehicle-specific procedures and requirements. As an example let it be mentioned here that in contrast to many other applications of face recognition, there is no administrator available in the vehicle for introduction to the process of enrollment and to provide help if recognition does not work. In addition, the recognition process in the vehicle should proceed completely transparently. This is an advantage of face recognition in particular, which can be carried out without contact and without special driver interaction. Various use cases and scenarios are presented and discussed. One approach used and the recognition rates achieved with it are presented.

The paper concludes with a forecast of existing synergies with other applications for an interior camera in the vehicle, such as drowsiness detection or recognition of gaze direction.