Detection of Facial Expressions Based on Three Feature Points Using Image Processing with Artificial Neural Networks

Previous attempts at Facial Emotion Detection have utilized both Eyes and Mouth when it comes to processing the image to feed into a classifier. These studies only consider these two critical features regarding facial detection. This study focuses more on using three features instead, adding the Eyebrows to the merged images created for the dataset. The study shows that further segmentation of key features in feature extraction allows for an increase in the average detection accuracy when classifying a given subject’s facial emotions. Using Artificial Neural Networks with a Backpropagation algorithm for training and testing emotion detection, the study aims to answer whether further segmentation can increase the accuracy of facial emotion detection and if it can be as accurate as previous studies. The resulting three-feature segmentation has increased accuracy in comparison to two-feature segmentation. Improvements can be made in the model, and the dataset used in the study is limited to the open-sourced datasets it has acquired for use