Deep Learning-Based Emotion Recognition from Facial Expressions

Abstract

Artificial Intelligence, Human-Computer Interaction and Human Intelligence are some of the most relevant areas of study at the present time. Especially the facial emotion recognition system that detects human emotions based on human expressions and human facial features has drawn many people's attention due to its wide usage in healthcare, security measures, advertisement, user behavior analysis, and user interfaces.
There are seven major emotions that are commonly distinguished and recognized in humans: happiness, sadness, anger, fear, surprise, disgust, and neutral expression. Facial emotion recognition is based on the changes in facial muscles and emotions' identification based on this feature. Many companies study facial emotions to understand their customers' response to certain goods and services offered and increase the quality of their services.
Random Forest and SVM algorithms have been successfully applied to machine learning. They work well with classification problems but convolutional neural networks are considered to be more accurate and applicable to emotion recognition due to their ability to identify human facial features.
Due to the development in deep learning techniques, modern facial recognition models based on CNNs are able to detect facial expressions in both images and video streams.
In this project, the facial emotion recognition system will be implemented with the use of CNN algorithm in Python with the help of the OpenCV library. The system will be developed based on the CNN classification algorithm for detecting facial features and emotions expressed with the help of facial images.