EUVIMEDEUVIMED entdecken
Crossref

Text-to-3D Scene Generation Using Generative AI

Donghui Li

Applied and Computational Engineering · 23.06.2026 · 242 (1) · 233-242

DOI: 10.54254/2755-2721/2026.gu34603 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied and Computational Engineering, online erschienen am 23.06.2026. Verfasst von Donghui Li.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The text-to-3D scene generation is one of the main directions in the study of generative artificial intelligence. It involves natural language processing, visual synthesis, geometric modeling and interactive virtual environments. Conventional development of 3D content usually demands specific modeling capabilities and high-cost software packages, as well as labor-intensive manual procedures and substantial time expenditure. Conversely, the new generative AI methods lower the costs of conventional 3D creation by enabling the user to just specify the object, room or environment that he/she wants in natural language. User instructions can be transformed into 3D objects, textured models, lighting or whole virtual scenes. A text-to-3D system enables a user to describe a scene such as a small bedroom with a desk, a bed and warm light and the AI can create a 3D version of it. This paper presents some of the most important techniques of text-to-3D generation such as diffusion models, neural radiance fields, 3D Gaussian point painting, mesh generation, and scene planning using large language models. This paper also mention the uses of these approaches in virtual reality, gaming, education and design. Despite the fact that the existing technology of 3D generation has enormous potential, there are many challenges such as lack of spatial consistency, unrealistic positioning of objects, lack of control over the users, low quality of assets and evaluation difficulties. The future text-to-3D systems could integrate generative AI, 3D engines, physics simulation, and manual editing to develop more realistic and practical virtual worlds.

Zitieren

APA

Li, D. (2026). Text-to-3D Scene Generation Using Generative AI. Applied and Computational Engineering, 242(1), 233-242. https://doi.org/10.54254/2755-2721/2026.gu34603

Vancouver

Li D. Text-to-3D Scene Generation Using Generative AI. Applied and Computational Engineering. 2026;242(1):233-242. doi:10.54254/2755-2721/2026.gu34603

BibTeX
@article{Li2026,
  author = {Li, Donghui},
  title = {Text-to-3D Scene Generation Using Generative AI},
  journal = {Applied and Computational Engineering},
  year = {2026},
  volume = {242},
  number = {1},
  pages = {233--242},
  doi = {10.54254/2755-2721/2026.gu34603}
}
Kostenlos starten →