Pushing the frontiers of unconstrained face detection and recognition: IARPA Janus Benchmark A

By Brendan Klare, Benjamin Klein, Emma Taborsky, Austin Blanton, J. Cheney, Kristen C. Allen, Patrick Grother, Alan Mah, M. Burge, Anil K. Jain. 2015

720 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Diversity in Faces (2019, methodology): “The gender annotations used two class labels (male and female) and the age group labeling used seven classes ([0-3],[4-12],[13-19],[20-30],[31-45],[46-60],[61-]), as well as a continuous age value to be consistent with the automatic prediction labels.”
VGGFace2: A Dataset for Recognising Faces across Pose and Age (2017, methodology): “Section V reports the results on the VGGFace2 test set, and state-of-the-art performance of several different architectures on the IJB-A [13] and IJBB [22] benchmarks.”
MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition (2016, methodology): “Available people images IJB-A [17] public 500 5712 LFW [7,8] public 5K 13K YFD [14] public 1595 3425 videos CelebFaces [15] public 10K 202K CASIA-WebFace [16] As shown in Table 1, our training dataset is considerably larger than the publicly available datasets.”
Continual Representation Learning for Biometric Identification (2020, background): “The recent improvements in biometric identiﬁcation are achieved by learning better representation with different losses, e.g. , softmax-based losses [11, 34, 67], triplet-based losses [9,20,52] and other kinds of losses [8,60], on large-scale image datasets [17,22,28,30,66,76,78].”
Comparator Networks (2018, background): “IJB-B Dataset [13] The IJB-B dataset is an extension of IJB-A [35], having 1, 845 subjects with 21.”
Two-branch Recurrent Network for Isolating Deepfakes in Videos (2020, background): “Unlike the proliferation of face recognition datasets [26,32,29,5,22], there has been a lack of large-scale face forensics datasets in the community for both training and evaluation.”
SSDMNV2: A real time DNN-based face mask detection system using single shot multibox detector and MobileNetV2 (2020, neutral): “Earlier datasets consisted of images fetched in supervised surroundings, while recent datasets are constructed by taking online images like WiderFace (Yang, Luo, Loy, & Tang, 2016), IJB-A (Klare et al., 2015a), MALF (Yang, Yan, Lei, & Li, 2015), and CelebA (Klare et al., 2015b).”
L2-constrained Softmax Loss for Discriminative Face Verification (2017, methodology): “We compare our algorithm with recently reported face veriﬁcation methods on LFW [14], YouTube Face [34] and IJB-A [16] datasets.”
LOTS about attacking deep features (2016, methodology): “[1] provide the current state-of-the-art results on the IJB-A benchmark [9], which is outperformed by the (unpublished) DNN of Ranjan et al.”
Feature Transfer Learning for Face Recognition With Under-Represented Data (2019, methodology): “Face recognition is one of the ongoing success stories in the deep learning era, yielding very high accuracy on several benchmarks [12,20,21].”
Gender shades : intersectional phenotypic and demographic evaluation of face datasets and gender classifiers (2017, methodology): “In response to these limitations IARPA, released the IJB-A dataset as the most geographically diverse set of collected faces, and no face detector was used to select images to limit bias (Klare et al., 2015).”
Feature Transfer Learning for Deep Face Recognition with Under-Represented Data (2018, methodology): “h non-identity variations such as pose and lighting. 1. Introduction Face recognition is one of the ongoing success stories in the deep learning era, yielding very high accuracy on several benchmarks [12,20,21]. However, it remains undetermined how deep learning classiﬁers for ﬁne-grained recognitio”
Disentangled Representation Learning GAN for Pose-Invariant Face Recognition (2017, methodology): “We believe that this is the first time such a fusing capability has been demonstrated on challenging databases such as IJB-A.”
Deep Domain Adaptation (2018, background): “Technical report, Technical Report 07-49, University of Massachusetts, Amherst, 2007 [22] B.”
UMDFaces: An annotated face dataset for training deep networks (2016, methodology): “We observe that the protocol used here is a tougher evaluation criterion than existing ones like LFW [10] and IJB-A [13].”