Relative Uncertainty Learning for Facial Expression Recognition

NeurIPS 2021  ·  Yuhang Zhang, Chengrui Wang, Weihong Deng ·

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.

PDF Abstract NeurIPS 2021 PDF NeurIPS 2021 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Facial Expression Recognition RAF-DB RUL (ResNet-18) Overall Accuracy 88.98 # 2

Methods


No methods listed for this paper. Add relevant methods here