Deep Expectation of Real and Apparent Age from a Single Image Without Facial Landmarks

By Rasmus Rothe, Radu Timofte, Luc Van Gool. 2016

938 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Public Health, Environmental and Occupational Health

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 Public Health, Environmental and Occupational Health

Citations in the corpus

Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing (2020, methodology): “oday, and none that enables an intersectional black box audit evaluation for such a wide range of tasks. To address this gap, we develop a benchmark called CelebSET, a subset of the IMDB-WIKI dataset [40] that includes 80 celebrity identities–20 of the most photographed celebrities from each of the ”
FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age (2019, background): “Many existing face datasets have relied on photographs of public ﬁgures such as politicians or celebrities [31, 23, 24, 46, 37].”
Diversity in Faces (2019, methodology): “We adopt the DEX model [8, 74] that is among the highest performing on some of the known face image data sets.”
Large image datasets: A pyrrhic win for computer vision? (2020, methodology): “We used the DEX [69] and the InsightFace [38] pre-trained models 15 16 and NSFW content ﬂagging information from the pre-trained NSFW-MobileNet-v2 model [34] to help perform a guided search of misogynistic consent-violating transgressions.”
Fully-Adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classification (2016, methodology): “The ﬁrst baseline is a VGG-16 model initialized from the a model trained from imdb-wiki gender classiﬁcation [30].”
Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification (2018, background): “…perception tasks including, but not limited to, face detection (Zafeiriou et al., 2015; Mathias et al., 2014; Bai and Ghanem, 2017), face classiﬁcation (Reid et al., 2013; Levi and Hassner, 2015a; Rothe et al., 2016) and face recognition (Parkhi et al., 2015; Wen et al., 2016; Ranjan et al., 2017).”
Third Time's the Charm? Image and Video Editing with StyleGAN3 (2022, methodology): “[57,58] for age, and the classiﬁer from Lin et al .”
AgeDB: The First Manually Collected, In-the-Wild Age Database (2017, background): “IMDB-WIKI [39, 40] 2015 523,051 20,284 Accurate to the year No Yes”
Auditing ImageNet: Towards a Model-driven Framework for Annotating Demographic Attributes of Large-Scale Image Datasets (2019, methodology): “In this work, we employ the Deep EXpectation (DEX) model of apparent age [22], which is pre-trained on the IMDB-WIKI dataset of 500k faces with real ages and ﬁne-tuned on the APPA-REAL training set of 2,476 images with apparent ages, crowdsourced from an average of 38 votes per image [1].”
A Comprehensive Analysis of Deep Regression (2018, neutral): “Regression could also be formulated as classiﬁcation [25], [26].”
Improving Smiling Detection with Race and Gender Diversity (2017, methodology): “To train the multihead (race and gender) diversity classifier, we use data collected from publicly available resources of celebrities appearing in [1, 33, 38], and further curate the data so that train and validation sets consist of disjoint sets of identities.”
Heterogeneous Face Attribute Estimation: A Deep Multi-Task Learning Approach (2017, methodology): “Compared with the best method on LAPAge2015 (DEX [17]), the proposed approach is about 10 times faster than a single VGG-16 model used in [17].”
CU-Net: Component Unmixing Network for Textile Fiber Identification (2019, methodology): “Recently, some CNN-based techniques (Wang et al. 2016; Wu et al. 2015b; Izadinia et al. 2015; Goutsu et al. 2018; Rothe et al. 2016) have been introduced for multi-label classification.”
FSA-Net: Learning Fine-Grained Structure Aggregation for Head Pose Estimation From a Single Image (2019, methodology): “Inspired by DEX [33], SSR-Net casts the regression problem of age estimation as a classification problem by dividing into the age domain into several age classes (bins).”
Understanding Unequal Gender Classification Accuracy from Face Images (2018, methodology): “Commonly used training datasets such as CelebA [21] and IMDb face [29] are made up of celebrities and biased towards light-skinned people.”
How to Train Your Energy-Based Model for Regression (2020, background): “Regression is a fundamental problem within computer vision with many additional applications [25, 42, 45, 54, 56], which all would benefit from such accurate predictions.”
Delving into Deep Imbalanced Regression (2021, neutral): “, 2009), which is a 100-class classification dataset, and (2) the IMDB-WIKI dataset (Rothe et al., 2018), which is a large-scale image dataset for age estimation from visual appearance.”
FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation (2021, methodology): “Many existing face datasets have been sourced from photographs of public figures such as politicians or celebrities [31, 21, 23, 48, 37].”