Development and application of face recognition technologies for automotive systems

This research involves the development and demonstration of a prototype automobile security system based on a face recognition system, which provides a highly reliable user verification in a non-intrusive system usable over a wide user population in a highly dynamic environment. Additional research demonstrated the possibility of developing systems for sleep on-set detection and DWI detection using the same input sensors an image processing technologies. The objectives of this research were to: (a) Identify research and development needed to make face recognition systems practical for implementation in automobiles. (b) Develop and implement new algorithms for face tracking and filtering to improve the performance of face recognition systems. (c) Develop and test different face databases to determine their effect on system performance and then to optimize performance using the best database. (d) Demonstrate and test a face recognition system in an automobile test bed and in a dynamic simulation. (e) Investigate the application of face recognition algorithms and technologies for the development of sleep onset detection in the automobile environment. (f) Develop and demonstrate a prototype sleep onset detection system in an auto-based simulator. (g) Investigate the use of face recognition algorithms and technologies in the development of an optical based DWI detection system. All of the objectives of this research were met, and a workable auto application of face recognition technology was demonstrated.