Multimodal sentiment analysis: A survey

Multimodal sentiment analysis has emerged as a prominent research field within artificial intelligence, benefiting immensely from recent advancements in deep learning. This technology has unlocked unprecedented possibilities for application and research, rendering it a highly sought-after area of study. In this review, we aim to present a comprehensive overview of multimodal sentiment analysis by delving into its definition, historical context, and evolutionary trajectory. Furthermore, we explore recent datasets and state-of-the-art models, with a particular focus on the challenges encountered and the future prospects that lie ahead. By offering constructive suggestions for promising research directions and the development of more effective multimodal sentiment analysis models, this review intends to provide valuable guidance to researchers in this dynamic field.