Graph neural network-based multimodal emotion recognition

Nowadays, Machine Learning (ML) and Deep Learning (DL) have been used to extract knowledge from human feelings. One of the objectives of this study is to use graph neural network (GNN) based emotion computing to analyze visual, speech, and textual data. Recognizing emotions is difficult because they change based on importance, age, culture, facial expression, surroundings scenario, and physical tiredness. For instance, physical exhaustion and psychological fatigue are two types of human feelings with significant differences. Existing studies have yet to completely include the most recent analysis developments and achievements in GNN-based emotion computing, which is one of our study's objectives. GNN is more acceptable for supporting realizing end-to-end sentiment analysis from visual, speech, and textual data. Rather than traditional emotion analysis techniques, our study relies on the process to achieve valuable accuracy and more specific results. Also, a study must be completed to bridge the gap between humans and machines.