A Survey on Semantic Communications System Based on Multi-modal Data

Abstract

The recent integration of deep learning with Semantic Communications (SC) has led to innovative advancements in the field of communication. In this survey paper, we extensively illuminate the research trends in SC systems that utilize deep learning, covering various single-modal data sources such as texts, images and videos, as well as multi-modal data sources. Through this, we provide a comprehensive understanding of how deep learning-based SC systems are rapidly advancing and how the current states and prospects of this technology are shaping up. This paper offers essential knowledge and information for researchers to grasp the technological developments and significance in this field and explore new research directions.