查看文章

This paper proposes new framework of communication system leveraging promising generation capabilities of multimodal generative models. Regarding nowadays’ smart applications, successful communication can be made by conveying the perceptual meaning, which we set as text prompt. Text serves as a suitable semantic representation of image data as it has evolved to instruct an image or generate image through mutli-modal techniques, by being interpreted in a manner similar to human cogitation. Utilizing text can also reduce the overload compared to transmitting the intact data itself. The transmitter converts objective image to text through multi-model generation process and the receiver reconstructs the image using reverse process. Each word in the text sentence has each syntactic role, responsible for particular piece of information the text contains. For further efficiency in communication load, the …