Deep Learning Based Facial Emotion Recognition using Multiple Layers Model

Abstract

A Facial Emotion Recognition (FER) system is an important tool to be implemented in any psychology academic field and beyond. This paper aims to show a system of Facial Emotion Recognition that can be done using the multiple layers model of ResN et50 which is an ANN. The dataset is an array that consists of 7 different emotions represented by the numbers 0–6. The emotion included in the dataset is "anger", "disgust", "fear", "happiness", "sadness", "shock", and "neutrality". This research found that the highest accuracy in the application of the model was recorded at 65% and the test was 60% to recognize facial emotion from a graphical input.