HUMAN EMOTION CLASSIFICATION USING DEEP LEARNING

Abstract

Research on facial detection and 
identification has received a lot of attention lately. 
Here, identifying and authenticating face traits is the 
primary goal of facial recognition. We have opted to 
investigate textual, audio, and visual inputs and 
create an ensemble model that compiles the data 
from all of these sources and presents it in a 
comprehensible and understandable manner. This 
approach can distinguish between seven different 
emotions: happiness, sadness, anger, surprise, fear, 
disgust, and neutrality. The three primary steps of 
the algorithm are feature extraction, emotion 
recognition, and image/audio/text processing. In this 
study, we used the algorithms CNN for recognizing 
emotions in video, SVM, HMM, and CNN for 
recognizing emotions in audio, and RNN, LSTM for 
recognizing emotions in text.