Data Uncertainty Learning in Face Recognition

By Jie Chang, Zhonghao Lan, Changmao Cheng, Yichen Wei. Computer Vision and Pattern Recognition, 2020

340 citations recorded by the source corpus.

Modeling data uncertainty is important for noisy images, but seldom explored for face recognition. The pioneer work, PFE, considers uncertainty by modeling each face image embedding as a Gaussian distribution. It is quite effective. However, it uses fixed feature (mean of the Gaussian) from an existing model. It only estimates the variance and relies on an ad-hoc and costly metric. Thus, it is not easy to use. It is unclear how uncertainty affects feature learning. This work applies data uncertainty learning to face recognition, such that the feature (mean) and uncertainty (variance) are learnt simultaneously, for the first time. Two learning methods are proposed. They are easy to use and outperform existing deterministic methods as well as PFE on challenging unconstrained scenarios. We also provide insightful analysis on how incorporating uncertainty estimation helps reducing the adverse effects of noisy samples and affects the feature learning.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/d0d955edbc44067e7fb469d1884eb59236dc770b

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Uncertainty-Aware Learning for Zero-Shot Semantic Segmentation (2020, methodology): “Besides serving as prediction conﬁdence as in the above methods, uncertainty is also modeled in distributional feature representation for applications like person ReID [58] and facial recognition [5, 49] to achieve robust feature learning.”
Probabilistic Embeddings for Cross-Modal Retrieval (2021, background): “[37] proposed the Hedged Instance Embedding (HIB) to handle the one-to-many correspondences for metric learning, while other works apply probabilistic embeddings to face understanding [43, 3], 2D-to-3D pose estimation [46], speaker diarization [44], and prototype embeddings [42].”
AdaFace: Quality Adaptive Margin for Face Recognition (2022, methodology): “To perform quality aware fusion, probabilistic approaches have been proposed to predict uncertainty in FR representation [2, 17, 26, 29].”
MagFace: A Universal Representation for Face Recognition and Quality Assessment (2021, methodology): “3) At last, the uncertainty method (DUL) performs well on CFP-FP but yields more verification errors on AgeDB-30 when the proportion of unconsidered images is increased.”