IARPA Janus Benchmark - C: Face Dataset and Protocol

By Brianna Maze, Jocelyn C. Adams, James A. Duncan, Nathan D. Kalka, Tim Miller, Charles Otto, Anil K. Jain, W. T. Niggel, Janet Anderson, J. Cheney and 1 more. 2018

701 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. Mei Wang
2. Liang Zheng
3. Xiaogang Wang
4. Yi Yang
5. Weihong Deng
6. Dacheng Tao
7. Wanli Ouyang
8. Andrew Zisserman

Citations in the corpus

• Diversity in Faces (2019, neutral): “Age Group Data set 0-3 4-12 13-19 20-30 31-45 46-60 > 60 LFW [15] 1.0% 10.6% 25.4% 29.6% 33.4% IJB-C [28] 0.0% 0.0% 0.5% 16.2% 35.5% 35.1% 12.7% Pubfig [35] 1.0% 10.8% 55.5% 21.0% 11.7% CelebA [9] 77.8% 22.1% UTKface [32] 8.8% 6.5% 5.0% 33.6% 22.6% 13.4% 10.1% AgeDB [33] 0.1% 0.52% 2.7% 17.5% 31.8%…”
• VGGFace2: A Dataset for Recognising Faces across Pose and Age (2017, methodology): “t 0:747 0:840 0:910 0:960 0:746 0:018 0:842 0:022 0:912 0:017 0:949 0:010 0:962 0:007 VGGFace2 ft SENet 0:768 0:862 0:927 0:967 0:763 0:018 0:865 0:018 0:914 0:020 0:951 0:013 0:961 0:010 Maze et al. [14] 0:600 0:750 0:860 0:950 0:450 0:620 0:790 0:870 0:900 Table VIII: Performance evaluation on the”
• ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2018, unknown): “Results on IJB-B and IJB-C.”
• Learning Compatible Embeddings (2021, methodology): “We use the first 11: 1:1 verification TAR (%@FAR=1e-4) on the IJB-C dataset [26] with different compatible directions. half of identities to train the old model and the full dataset to train the new model.”
• ElasticFace: Elastic Margin Loss for Deep Face Recognition (2021, background): “Similarly, on the IJB-C benchmark, our ElasticFace-Cos scored a TAR at FAR1e–4 of 96.57%, ahead of the Partial-FC-CosFace [1] (pre-print) and the GroupFace [13] (CVPR2020) with 96.4% and 96.36% respectively.”
• Global-Local GCN: Large-Scale Label Noise Cleansing for Face Recognition (2020, methodology): “identities overlapping with LFW [26], FaceScrub [37] and IJB [34, 46].”
• ShrinkTeaNet: Million-scale Lightweight Face Recognition via Shrinking Teacher-Student Networks (2019, background): “IJB-B [48] and IJB-C [49] are introduced as two large-scale face verification protocols.”
• Data Uncertainty Learning in Face Recognition (2020, methodology): “2 benchmarks including LFW [18] and MegaFace [20] 5 , and 3 unconstrained benchmarks: CFP [34] 6 , YTF [43] and IJB-C [28], are used to evaluate the performance of DUL cls/rgs following the standard evaluation protocols.”
• Past, Present, and Future of Face Recognition: A Review (2020, unknown): “Database Apparition’s Date Images Subjects Images/Subject ORL [23] 1994 400 40 10 FERET [13] 1996 14,126 1199 - AR [24] 1998 3016 116 26 XM2VTS [25] 1999 - 295 - BANCA [26] 2003 - 208 - FRGC [14] 2006 50,000 - 7 LFW [10] 2007 13,233 5749 ≈2.3 CMU Multi PIE [29] 2009 >750,000 337 N/A IJB-A [31] 2015 ”
• Boundary-aware Backward-Compatible Representation via Adversarial Learning in Image Retrieval (2023, background): “Researchers pay more attention to the improvement of retrieval performance [3, 13, 25, 27, 28] and there exist several comprehensive benchmarks [12, 22].”
• FSGAN: Subject Agnostic Face Swapping and Reenactment (2019, methodology): “We use the video sequences of the IJB-C dataset [30] to train our generator, G r , for which we automatically extracted the frames depicting particular subjects.”
• Deep Face Recognition: A Survey (2018, background): “Three current major datasets, namely, MegaFace [44], [164] , MS-Celeb-1M [45] and IJB-A/B/C [41], [42], [43], are corresponding to largescale FR with a very large number of candidates, low/oneshot FR and large pose-variance FR which will be the focus of research in the future.”
• AdaCos: Adaptively Scaling Cosine Logits for Effectively Learning Deep Face Representations (2019, methodology): “To demonstrate the effectiveness of the proposed AdaCos loss function, we evaluated it on several face benchmarks, including LFW face verification [13], MegaFace one-million identification [24] and IJB-C [23].”
• Sub-center ArcFace: Boosting Face Recognition by Large-Scale Noisy Web Faces (2020, methodology): “Our testing datasets consist of IJB-B [35], IJB-C [17], MegaFace [13], and Face Recognition Vendor Test (FRVT).”
• Comparator Networks (2018, methodology): “In § 5, we report the verification performance of the DCN on the challenging IARPA Janus face recognition benchmarks – IJB-B [13] and IJB-C [14].”