Seven Basic Expression Recognition Using ResNet-18

TL;DR

This paper presents a ResNet-18 based approach for recognizing seven basic facial expressions in-the-wild, achieving improved accuracy over baseline models by employing data resampling, augmentation, and regularization techniques.

Contribution

The study demonstrates the effectiveness of using a pre-trained ResNet-18 architecture with data balancing and augmentation for affective expression recognition in unconstrained environments.

Findings

Achieved an ABAW2 score of 0.4, surpassing baseline results.

Implemented class-wise resampling and data augmentation to handle class imbalance.

Validated the approach on a large dataset with 2.8 million frames.

Abstract

We propose to use a ResNet-18 architecture that was pre-trained on the FER+ dataset for tackling the problem of affective behavior analysis in-the-wild (ABAW) for classification of the seven basic expressions, namely, neutral, anger, disgust, fear, happiness, sadness and surprise. As part of the second workshop and competition on affective behavior analysis in-the-wild (ABAW2), a database consisting of 564 videos with around 2.8M frames is provided along with labels for these seven basic expressions. We resampled the dataset to counter class-imbalances by under-sampling the over-represented classes and over-sampling the under-represented classes along with class-wise weights. To avoid overfitting we performed data-augmentation and used L2 regularisation. Our classifier reaches an ABAW2 score of 0.4 and therefore exceeds the baseline results provided by the hosts of the competition.