Related Experiment Video
Updated: Jun 14, 2026
19:32
Contextual and Cued Fear Conditioning Test Using a Video Analyzing System in Mice
Published on: March 1, 2014
FRVT 2006 and ICE 2006 large-scale experimental results
P Jonathon Phillips1, W Todd Scruggs, Alice J O'Toole
1National Institute of Standards and Technology (NIST), Gaithersburg, MD 20899, USA. jonathon@nist.gov
Summary
Large-scale tests show significant improvements in face and iris recognition technology. Algorithms now outperform humans in matching faces under varying illumination conditions.
Area of Science:
- Biometrics
- Computer Vision
- Pattern Recognition
Background:
- Biometric recognition systems are crucial for security and identification.
- Previous evaluations like FRVT 2002 established baseline performance metrics.
- Advancements in imaging and algorithms necessitate updated large-scale evaluations.
Purpose of the Study:
- To present large-scale experimental results from the Face Recognition Vendor Test (FRVT) 2006 and Iris Challenge Evaluation (ICE) 2006.
- To compare the recognition performance of high-resolution still frontal face, 3D face, and iris images.
- To assess algorithm performance against human accuracy in face recognition under challenging illumination.
Main Methods:
- Evaluated recognition performance on high-resolution still frontal face, 3D face, and iris images.
- Included images of varying quality, even those failing standard quality control, in ICE 2006.
- Conducted experiments comparing algorithm and human accuracy on face matching across different illumination conditions.
Main Results:
- FRVT 2006 demonstrated at least an order-of-magnitude improvement in recognition performance over FRVT 2002 for controlled still and 3D face images.
- Recognition performance was comparable across high-resolution frontal face, 3D face, and iris images on FRVT 2006 and ICE 2006 datasets.
- The best-performing algorithms exceeded human accuracy in matching face identity under changing illumination.
Conclusions:
- Significant progress has been made in face and iris recognition technologies.
- High-resolution frontal face, 3D face, and iris modalities offer comparable recognition performance.
- Automated systems show potential to surpass human capabilities in specific biometric identification tasks.
