Its About the Face Impostor Distribution

We studied the effects of factors on the false accept rate (FAR) for three modern video face recognition algorithms. We examined the effects of environment (location), video(imagery-) based, and demographic factors. The study is performed on the handheld video in the Point and Shoot Face Recognition Challenge (PaSC), which consists of 1401 handheld videos of 265 subjects. The results of our analysis are consistent across the three algorithms. Our analysis shows that FAR can significantly vary. Surprisingly, for environment and video-based factors, there was a clear relationship between verification rate (VR) and FAR. An increase (resp. decrease) in the FAR results in an increase (resp. decrease) in the VR. We looked at the shape of the marginal impostor distributions for each level of a factor. In most cases these impostor distributions for a given algorithm moved according to a simple affine transform, translation and scaling, when moving between factor levels.

[1]  Alice J. O'Toole,et al.  Demographic effects on estimates of automatic face recognition performance , 2011, Face and Gesture 2011.

[2]  Bruce A. Draper,et al.  Biometric zoos: Theory and experimental evidence , 2011, 2011 International Joint Conference on Biometrics (IJCB).

[3]  Patrick J. Flynn,et al.  SNoW: Understanding the causes of strong, neutral, and weak face impostor pairs , 2013, 2013 IEEE Sixth International Conference on Biometrics: Theory, Applications and Systems (BTAS).

[4]  Gang Hua,et al.  Eigen-PEP for Video Face Recognition , 2014, ACCV.

[5]  Bruce A. Draper,et al.  Focus on quality, predicting FRVT 2006 performance , 2008, 2008 8th IEEE International Conference on Automatic Face & Gesture Recognition.

[6]  Bruce A. Draper,et al.  Computational Statistics and Data Analysis Introduction to Face Recognition and Evaluation of Algorithm Performance , 2022 .

[7]  James J. Filliben,et al.  Generalizing face quality and factor measures to video , 2014, IEEE International Joint Conference on Biometrics.

[8]  Takeo Kanade,et al.  Multi-PIE , 2008, 2008 8th IEEE International Conference on Automatic Face & Gesture Recognition.

[9]  Bruce A. Draper,et al.  Report on the FG 2015 Video Person Recognition Evaluation , 2015, 2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG).

[10]  Neil Yager,et al.  The Biometric Menagerie , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Bruce A. Draper,et al.  An introduction to the good, the bad, & the ugly face recognition challenge problem , 2011, Face and Gesture 2011.

[12]  Gang Hua,et al.  Probabilistic Elastic Matching for Pose Variant Face Verification , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[13]  Bruce A. Draper,et al.  The challenge of face recognition from digital point-and-shoot cameras , 2013, 2013 IEEE Sixth International Conference on Biometrics: Theory, Applications and Systems (BTAS).

[14]  Shiguang Shan,et al.  Hybrid Euclidean-and-Riemannian Metric Learning for Image Set Classification , 2014, ACCV.

[15]  Simon Dobrisek,et al.  Modest face recognition , 2015, 3rd International Workshop on Biometrics and Forensics (IWBF 2015).

[16]  Alice J. O'Toole,et al.  Demographic effects on estimates of automatic face recognition performance , 2012, Image Vis. Comput..