Report on the Evaluation of 2D Still-Image Face Recognition Algorithms

By Patrick Grother, G. W. Quinn, P. Jonathon Phillips. 2011

212 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Xiaogang Wang
2. Dacheng Tao
3. Percy Liang
4. Yoshua Bengio
5. Andrew Zisserman
6. Yi Yang
7. Weihong Deng
8. Mei Wang

Citations in the corpus

• Fairness Without Demographics in Repeated Loss Minimization (2018, background): “This representation disparity forms our definition of unfairness, and has been observed in face recognition (Grother et al., 2011), language identification (Blodgett et al.”
• The MegaFace Benchmark: 1 Million Faces for Recognition at Scale (2015, background): “Large scale evaluations were previously performed on controlled datasets (visa photographs, mugshots, lab captured photos) by NIST [12], and report recognition results of 90% on 1.6 million people.”
• Web-scale training for face identification (2014, methodology): “The datasets used by [7, 11] are not publicly available, and therefore it is hard to compare the performance of academic systems on the same benchmarks.”
• Unconstrained Face Recognition: Identifying a Person of Interest From a Media Collection (2014, methodology): “The state-of-the-art COTS face matcher used in our experiments was one of the top performers in the 2010 NIST Multi-Biometric Evaluation (MBE) [38].”
• Racial Faces in the Wild: Reducing Racial Bias by Information Maximization Adaptation Network (2018, background): “Several studies [49, 29, 23, 50, 36] have uncovered racial bias in non-deep 1 http://www.whdeng.cn/RFW/index.html face recognition algorithms.”
• MegaFace: A Million Faces for Recognition at Scale (2015, background): “Large scale evaluations were performed only on controlled datasets (visa photographs, mugshots, lab captured photos) by NIST [10], and report recognition results of 90% on 1.6 million people.”
• Face Recognition Vendor Test (FRVT) Performance of Face Identification Algorithms NIST IR 8009 (2014, methodology): “In any case, the 2010 NIST evaluation of face recognition showed that considerable accuracy benefits accrue with retention and use of all historical images [12].”
• Partial Face Recognition: Alignment-Free Approach (2013, background): “2010 [2], face recognition in controlled conditions (frontal face of cooperative users and controlled indoor illumination) has already achieved impressive performance over large-scale galleries.”
• Deep Face Recognition: A Survey (2018, background): “According to [109], [64], [17], the female, Black, and younger cohorts are usually more difficult to recognize for non-deep FR algorithms due to this data bias.”
• Learning Models with Uniform Performance via Distributionally Robust Optimization (2018, background): “In numerous other applications—such as facial recognition, automatic video captioning, language identification, academic recommender systems— performance varies significantly over different demographic groupings, such as race, gender, or age [46, 50, 20, 80, 91].”
• Learning Race from Face: A Survey (2014, background): “In particular, experimental results in [28], [29], [30], [31], [68], [169], [175], [220], [221], [222] have shown that both commercial and nontrainable face recognition algorithms consistently tend to perform lower on the same cohorts (females, African”