• The Course
• The Lecturers
• School venue
• Travel and accomodation
• Pictures from 2005

Download:

• Application form
• Registration form

Biological recognition of human faces

Prof. Alice O'Toole
University of Texas at Dallas

The accuracy of computational models of face recognition has been tested extensively over the last decade. Virtually nothing is known, however, about the accuracy of face recognition algorithms relative to humans.

In the last two years, we have begun a series of direct comparisons between state-of-the-art face recognition algorithms, being tested in the U.S. Government-sponsored “Face Recognition Grand Challenge” with human performance on the same task.

In this presentation, I will discuss the methods we have used in these comparisons, the lessons learned, and the results to date. I will also discuss the challenges to sampling the enormous amounts of data available from algorithms for making useful and valid comparisons to human memory and perception.

The evaluation of algorithms relative to humans provides insight into both the pitfalls and advantages of the human system relative to computer-based algorithms and is informative for the development of hybrid systems that use algorithms and humans to their best advantage.

with support from