Proceedings of the 2nd International Conference on Art Design and Digital Technology, ADDT 2023, September 15–17, 2023, Xi’an, China

Research Article

Behind the AI Art Creation: A Study of Generative Models for Text-to-Image Generation

Download56 downloads
  • @INPROCEEDINGS{10.4108/eai.15-9-2023.2340842,
        author={Lege  Zhao and Han  Zhang},
        title={Behind the AI Art Creation: A Study of Generative Models for Text-to-Image Generation},
        proceedings={Proceedings of the 2nd International Conference on Art Design and Digital Technology, ADDT 2023, September 15--17, 2023, Xi’an, China},
        publisher={EAI},
        proceedings_a={ADDT},
        year={2024},
        month={1},
        keywords={ai art creation; test-to-image generation; generative models},
        doi={10.4108/eai.15-9-2023.2340842}
    }
    
  • Lege Zhao
    Han Zhang
    Year: 2024
    Behind the AI Art Creation: A Study of Generative Models for Text-to-Image Generation
    ADDT
    EAI
    DOI: 10.4108/eai.15-9-2023.2340842
Lege Zhao1, Han Zhang1,*
  • 1: Dongbei University of Finance and Economics
*Contact email: hanzhang@dufe.ed.cn

Abstract

The advancement of deep learning has greatly facilitated computer vision and natural language processing. Among its applications is text-to-image generation, which involves creating images from textual descriptions. Recent text-to-image techniques offer a compelling yet straightforward ability to convert text into images, making them a prominent research topic in both AI and art creation. Image generation from text holds a myriad of practical and creative applications in computer design and the creation of digital art. This paper conducts a comprehensive study to review three types of generative models for text-to-image generation, aiming to provide a foundational understanding of the principles underlying these models.