Finding Emotions in Faces: A Meta-Classifier

TL;DR

This paper introduces a meta-classifier that combines facial landmark-based features and deep learning to improve emotion recognition accuracy in faces from 58% to 77%.

Contribution

It presents a novel meta-classifier approach that integrates two different emotion recognition methods for enhanced performance.

Findings

Meta-classifier achieves 77% accuracy.

Combining methods outperforms individual approaches.

Different methods complement each other on varied images.

Abstract

Machine learning has been used to recognize emotions in faces, typically by looking for 8 different emotional states (neutral, happy, sad, surprise, fear, disgust, anger and contempt). We consider two approaches: feature recognition based on facial landmarks and deep learning on all pixels; each produced 58% overall accuracy. However, they produced different results on different images and thus we propose a new meta-classifier combining these approaches. It produces far better results with 77% accuracy