Emotion Recognition in the Wild using Deep Neural Networks and Bayesian Classifiers

TL;DR

This paper introduces a hybrid approach combining deep neural networks and Bayesian classifiers for group emotion recognition in unconstrained environments, achieving significant accuracy improvements over baseline methods.

Contribution

It presents a novel combination of bottom-up neural analysis and top-down Bayesian classification for emotion recognition in challenging real-world images.

Findings

Achieved 64.68% accuracy on the ERW Challenge dataset.

Outperformed the baseline accuracy of 53.62%.

Validated the effectiveness of the hybrid model in wild conditions.

Abstract

Group emotion recognition in the wild is a challenging problem, due to the unstructured environments in which everyday life pictures are taken. Some of the obstacles for an effective classification are occlusions, variable lighting conditions, and image quality. In this work we present a solution based on a novel combination of deep neural networks and Bayesian classifiers. The neural network works on a bottom-up approach, analyzing emotions expressed by isolated faces. The Bayesian classifier estimates a global emotion integrating top-down features obtained through a scene descriptor. In order to validate the system we tested the framework on the dataset released for the Emotion Recognition in the Wild Challenge 2017. Our method achieved an accuracy of 64.68% on the test set, significantly outperforming the 53.62% competition baseline.