Add to Chrome
Add to Firefox
Add to EdgeLeveraging Synthetic Data for Generalizable and Fair Facial Action Unit Detection
Mar 15, 2024
Facial action unit (AU) detection is a fundamental block for objective facial expression analysis. Supervised learning approaches require a large amount of manual labeling which is costly. The limited labeled data are also not diverse in terms of gender which can affect model fairness. In this paper, we propose to use synthetically generated data and multi-source domain adaptation (MSDA) to address the problems of the scarcity of labeled data and the diversity of subjects. Specifically, we propose to generate a diverse dataset through synthetic facial expression re-targeting by transferring the expressions from real faces to synthetic avatars. Then, we use MSDA to transfer the AU detection knowledge from a real dataset and the synthetic dataset to a target dataset. Instead of aligning the overall distributions of different domains, we propose Paired Moment Matching (PM2) to align the features of the paired real and synthetic data with the same facial expression. To further improve gender fairness, PM2 matches the features of the real data with a female and a male synthetic image. Our results indicate that synthetic data and the proposed model improve both AU detection performance and fairness across genders, demonstrating its potential to solve AU detection in-the-wild.
