.png){kind=icon}
.png){kind=icon}
Return
DuaFace: Data uncertainty in angular based loss for face recognition
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patrec.2023.01.013.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The Data Uncertainty inherently existed in feature continuous mapping space itself and the training dataset. In this paper, a general loss function DuaFace based on Data Uncertainty and Angular/cosine-margin-based loss is proposed to study the influence of Data Uncertainty on traditional Angular based loss functions. Correspondingly, insightful analysis on how incorporating Data Uncertainty estimation helps reducing the adverse effects of noisy samples and affects the process of feature learning are also provided. Moreover, extensive experiments conducted on Face Recognition demonstrate its superiority over state-of-the-arts. (c) 2023 Elsevier B.V. All rights reserved.
Keywords:
Face recognition
Loss function
Data uncertainty