Relative Uncertainty Learning for Facial Expression Recognition

Yuhang Zhang, Chengrui Wang, Weihong Deng

Keywords: uncertainty learning, facial expression recognition, label noise

TL;DR: Our model learns uncertainty from the relativity of different samples.

Abstract: In facial expression recognition (FER), the uncertainties introduced by inherent noises like ambiguous facial expressions and inconsistent labels raise concerns about the credibility of recognition results. To quantify these uncertainties and achieve good performance under noisy data, we regard uncertainty as a relative concept and propose an innovative uncertainty learning method called Relative Uncertainty Learning (RUL). Rather than assuming Gaussian uncertainty distributions for all datasets, RUL builds an extra branch to learn uncertainty from the relative difficulty of samples by feature mixup. Specifically, we use uncertainties as weights to mix facial features and design an add-up loss to encourage uncertainty learning. It is easy to implement and adds little or no extra computation overhead. Extensive experiments show that RUL outperforms state-of-the-art FER uncertainty learning methods in both real-world and synthetic noisy FER datasets. Besides, RUL also works well on other datasets such as CIFAR and Tiny ImageNet. The code is available at https://github.com/zyh-uaiaaaa/Relative-Uncertainty-Learning.

Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.