DeepFN: Towards Generalizable Facial Action Unit Recognition with Deep Face Normalization
This work addresses the problem of deploying facial action recognition systems in real-world applications like market research and psychotherapy by improving generalization across individuals, genders, skin types, and datasets, though it is incremental as it builds on existing normalization and self-supervised learning approaches.
The paper tackled the limited generalization of facial action unit recognition models across unseen people and demographics by introducing DeepFN, a deep face normalization method that transfers expressions onto a common template, reducing generalization gaps: for person-independent models, it increased performance from 55% to 59.6% F1/accuracy, cutting the gap from 5.3% to near-zero, and similarly reduced gaps for gender (2.4%), skin type (5.3%), and dataset (9.4%).
Facial action unit recognition has many applications from market research to psychotherapy and from image captioning to entertainment. Despite its recent progress, deployment of these models has been impeded due to their limited generalization to unseen people and demographics. This work conducts an in-depth analysis of performance across several dimensions: individuals(40 subjects), genders (male and female), skin types (darker and lighter), and databases (BP4D and DISFA). To help suppress the variance in data, we use the notion of self-supervised denoising autoencoders to design a method for deep face normalization(DeepFN) that transfers facial expressions of different people onto a common facial template which is then used to train and evaluate facial action recognition models. We show that person-independent models yield significantly lower performance (55% average F1 and accuracy across 40 subjects) than person-dependent models (60.3%), leading to a generalization gap of 5.3%. However, normalizing the data with the newly introduced DeepFN significantly increased the performance of person-independent models (59.6%), effectively reducing the gap. Similarly, we observed generalization gaps when considering gender (2.4%), skin type (5.3%), and dataset (9.4%), which were significantly reduced with the use of DeepFN. These findings represent an important step towards the creation of more generalizable facial action unit recognition systems.