Research on Semantic Communication Based on Balancing of Task Distortion

Le résumé fourni par la source

With the development of the Artificial Intelligence (AI) and communication technology, how to improve communication efficiency for the semantic communication system in image reconstruction and image classification tasks, meanwhile how to balance these two types of tasks is a hot research topic in the field of semantic communication. This article proposes a semantic communication framework that balances task distortion based on Grad CAM technology, obtains the optimal semantic features that can balance reconstruction quality and classification accuracy, and further compresses the semantic features based on task relevance. Finally, the framework was used in the course design of “Digital Image Communication” and verified that the proposed algorithm can balance image reconstruction and image classification tasks while further reducing the amount of semantic data transmitted.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.