Image Transmission

Image transmission research focuses on efficiently and reliably sending images across various channels, prioritizing both fidelity and semantic meaning. Current efforts concentrate on deep learning-based joint source-channel coding (JSCC) methods, often employing autoencoders, variational autoencoders, diffusion models, and transformers to achieve high compression ratios and robust reconstruction even under noisy conditions. These advancements leverage semantic information extraction and generative models to improve perceptual quality and adaptability to different bandwidths and channel characteristics, impacting fields like 6G communication and multimedia applications. The development of new metrics for evaluating semantic similarity is also a key area of investigation.