Research Article
Enhanced Semantic Communication in 6G Networks Using DCGAN
@INPROCEEDINGS{10.1007/978-3-031-81168-5_12, author={Sowmya Sri Nalluri and G. A. E. Satish Kumar and Dileep Kumar Arumulla and Vinod Kumar Auti}, title={Enhanced Semantic Communication in 6G Networks Using DCGAN}, proceedings={Broadband Communications, Networks, and Systems. 14th EAI International Conference, BROADNETS 2024, Hyderabad, India, February 16--17, 2024, Proceedings, Part I}, proceedings_a={BROADNETS}, year={2025}, month={2}, keywords={Semantic Communications Deep Convolutional Generative Adversarial Network (DCGAN) Encoder and Decoder}, doi={10.1007/978-3-031-81168-5_12} }- Sowmya Sri Nalluri
G. A. E. Satish Kumar
Dileep Kumar Arumulla
Vinod Kumar Auti
Year: 2025
Enhanced Semantic Communication in 6G Networks Using DCGAN
BROADNETS
Springer
DOI: 10.1007/978-3-031-81168-5_12
Abstract
Semantic communication diverges from Shannon’s communication theory by prioritizing the semantic essence of data over its step-by-step reconstruction at the receiver’s end, signifying its potential to shape the future of mobile communication. This approach aims to address the limitations posed by finite bandwidth in transmitting information for modern, high-volume multimedia applications. Leveraging the integration of AI technology with 6G networks, it provides complete communication systems built on semantic communication concepts. This research focuses on creating an end-to-end picture transmission system based on semantic communication by investigating important design factors that are linked with physical channel features.
To achieve transmission of realistic images from semantically segmented inputs, previously trained DCGAN (Deep Convolutional Generative Adversarial Network) model is used at the target end., trained using COCO-Stuff dataset for both receiver DCGAN (decoder) and transmitter semantic segmentation (encoder). Notably, the study unveils that broadcasting semantic segmentation maps, rather than actual images, across the physical channel yields substantial resource gains, particularly in bandwidth conservation compared to conventional communication methods. Additionally, the research delves into examining the effects of quantization noise and physical channel irregularities on multimedia content transfer facilitated by semantic communication.
