Research Summary

Participants: Jinwoo Kang (Lead), Monson Hayes, Amol Borkar and Mark Smith

This research is concerned with the design and development of a consumer grade biometric system for illumination invariant face recognition that is intended to be used in the automotive market for vehicle personalization. The primary goal of the research that is described in this paper is to investigate the feasibility of practical face recognition used for identity management in order to minimize algorithmic complexity while making the system robust to ambient illumination changes.

Biometrics as an identity management discipline seeks to either identify a person, or verify a person's claimed identity. In most such systems a design goal is to minimize the probability of false admission even at the expense of increasing the probability of false rejection. The economics behind this are driven by applications that must manage a significant economic, safety or security threat. Any such system that incorrectly admits even a vanishingly small number of subjects will fail in the market place. However, the market for ID management extends well beyond high security applications. Proponents of pervasive and context aware computing have argued that a knowledge of the user's identity can greatly enhance the perceived value of an application by personalizing it.

In this research, face recognition is used because it is a well studied biometric using both static and video based image collection, and also has the attributes of being noninvasive and potentially requiring no explicit co-operation from the user for the biometric to work. With respect to performance, an interesting difference in this work compared with previous face recognition research is the idea of consumer grade biometrics, and how it can alter the goals of biometric device design.


Camera and LED Array Mounted in Car

Near IR LED Array


  1. J. Kang, A. Borkar, A. Yeung, N. Nong, M. Smith, and M. Hayes, ”Short wavelength infrared face recognition for personalization”, Proc. 2006 Int. Conf. on Image Processing, paper WA-P8.8, pp. 2759-2762, Oct. 2006.


Jinwoo Kang

Personal Web Page

Return to top