A Deep Learning Perspective on the Origin of Facial Expressions

TL;DR

This paper explores how convolutional neural networks understand facial expressions, linking their features to established systems like FACS, and demonstrates improved micro-expression detection using transfer learning and LSTM models.

Contribution

It provides insights into CNN feature representations related to FACS and AU, and introduces a novel approach for micro-expression detection with state-of-the-art accuracy.

Findings

CNN features correlate with FACS and AU

Transfer learning improves cross-dataset performance

LSTM enhances micro-expression detection accuracy

Abstract

Facial expressions play a significant role in human communication and behavior. Psychologists have long studied the relationship between facial expressions and emotions. Paul Ekman et al., devised the Facial Action Coding System (FACS) to taxonomize human facial expressions and model their behavior. The ability to recognize facial expressions automatically, enables novel applications in fields like human-computer interaction, social gaming, and psychological research. There has been a tremendously active research in this field, with several recent papers utilizing convolutional neural networks (CNN) for feature extraction and inference. In this paper, we employ CNN understanding methods to study the relation between the features these computational networks are using, the FACS and Action Units (AU). We verify our findings on the Extended Cohn-Kanade (CK+), NovaEmotions and FER2013 datasets. We apply these models to various tasks and tests using transfer learning, including cross-dataset validation and cross-task performance. Finally, we exploit the nature of the FER based CNN models for the detection of micro-expressions and achieve state-of-the-art accuracy using a simple long-short-term-memory (LSTM) recurrent neural network (RNN).