Download: |
Grand Challenges and Progress in Face Recognition and Visual BiometricsDr. P. Jonathon Phillips Over the last twenty years, phenomenal progress has been made in recognizing frontal faces taken in mobile studio and mugshot environments. From 1993 to 2010, the false reject rate at a false accept rate of 1 in 1000 has halved every two years. When the imaging constraints for are relaxed, performance degrades. The ubiquity of cell phone and digital point and shoot cameras makes face recognition from these cameras an important challenge. This includes face recognition in still images and videos. In this talk I will discuss the challenges of recognizing faces and people in point and shoot images and videos. The first set of challenges is designing algorithms that can process still images and videos. The second set of challenges includes measuring performance and setting achievable performance goals. |
|