AgeDB: The First Manually Collected, In-the-Wild Age Database

By Stylianos Moschoglou, Athanasios Papaioannou, Christos Sagonas, Jiankang Deng, Irene Kotsia, Stefanos P. Zafeiriou. 2017

877 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Computational Mechanics

In Computer Vision and Pattern Recognition

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.”
Discrimination-Aware Network Pruning for Deep Model Compression (2020, methodology): “2) Data sets and implementation details: We evaluate the proposed DCP method on four benchmark data sets, including LFW [30], CFP-FP [62], AgeDB-30 [54], and MegaFace [36].”
StyleGAN2 Distillation for Feed-forward Image Manipulation (2020, background): “CACD [8], AgeDB [40], although with different expressions and face orientation.”
Marginal Loss for Deep Face Recognition (2017, methodology): “The experiments were conducted on a subset of the final version of AgeDB, as AgeDB was further extended by the time it became publicly available [15].”
Quality-Agnostic Image Recognition via Invertible Decoder (2021, methodology): “Accuracy (%) on face benchmarks such as AgeDB-30 [34] and CFP-FP [39] in four resolution types.”
ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2018, background): “LFW [10] 5,749 13,233 CFP-FP [28] 500 7,000 AgeDB-30 [19] 568 16,488 CPLFW [44] 5,749 11,652 CALFW [45] 5,749 12,174 YTF [38] 1,595 3,425”
RetinaFace: Single-stage Dense Face Localisation in the Wild (2019, background): “LFW [24], CFP-FP [46], AgeDB-30 [35] and IJBC [34]), and keep the embedding network ( i.e .”
CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition (2020, methodology): “We extensively test our method on several popular benchmarks, including LFW [9], CFP-FP [24], CPLFW [41], AgeDB [20], CALFW [40], IJB-B [33], IJB-C [19], and MegaFace [11].”
WebFace260M: A Benchmark Unveiling the Power of Million-Scale Deep Face Recognition (2021, methodology): “Since public evaluations are most saturated [26, 37, 49] and may contain noise [29, 36], we manually construct a new test set with rich attributes to enable FRUITS, including different age, gender, race and scenario evaluations.”
The Shallow End: Empowering Shallower Deep-Convolutional Networks through Auxiliary Outputs (2016, methodology): “We adopt the large-scale data set MS1M [8] as the training data and four benchmark data sets as the validation data, including LFW [20], CFP-FP [38], AgeDB-30 [31], and MegaFace [21].”
Delving into Deep Imbalanced Regression (2021, neutral): “• AgeDB-DIR (age) : AgeDB-DIR is constructed in a similar manner from the AgeDB dataset (Moschoglou et al., 2017).”
Sub-center ArcFace: Boosting Face Recognition by Large-Scale Noisy Web Faces (2020, methodology): “Besides, we also report our ﬁnal results on widely used veriﬁcation datasets (e.g. LFW [12], CFP-FP [27], and AgeDB-30 [18]).”
SphereFace Revived: Unifying Hyperspherical Face Recognition (2021, neutral): “8M train LFW [77] 5,749 13,233 validation AgeDB-30 [78] 568 16,488 validation CALFW [79] 5,749 11,652 validation CPLFW [80] 5,749 12,174 validation CFP [81] 500 7,000 validation VGGFace2_test [75] 500 173k validation IJB-B [82] 1,845 76.”
Person Recognition in Personal Photo Collections (2017, background): “Database (AgeDB) [20] evaluates the recognition across long”
Exclusivity-Consistency Regularized Knowledge Distillation for Face Recognition (2020, methodology): “AgeDB [32] contains images annotated with accurate to the year, noise-free labels.”
ShrinkTeaNet: Million-scale Lightweight Face Recognition via Shrinking Teacher-Student Networks (2019, methodology): “AgeDB [47] provides a protocol with 4 testing age groups where each group consists of 10 splits of 600 matching pairs from 440 subjects.”