THE USE OF GENERATIVE NEURAL NETWORKS IN PROCEDURAL MODELING OF GAME WORLDS
The article explores the application of generative neural networks in procedural modeling of game worlds, enabling the automation of game content creation, including levels, scenarios, 3D objects, textures, and NPC behavior.The study aims to analyze the role of generative neural networks in shaping game environments, identify key types of generated content, and examine modern algorithms and models that enhance procedural generation.The study employed general scientific methods of cognition based on analysis, synthesis, comparison of scientific literature, and expert publications.The findings indicate that generative neural networks serve as a crucial tool in procedural content creation.They facilitate the efficient automation of game design development, significantly reducing the time required to construct game environments.The study identified key types of generated content, including graphical elements, audio components, spatial structures, interaction systems, storylines, balancing mechanics, and other supplementary data.The analysis revealed that different neural network architectures are applied to generate various aspects of the gaming experience.Specifically, convolutional neural networks are used for processing images and textures; recurrent neural networks generate scripts and event sequences; autoencoders compress and optimize data; while generative adversarial networks (GANs) create realistic objects and structures.Advanced technologies enhance 3D visualization and automated game object generation, contributing to the development of more realistic and logically structured gaming environments.The study suggests that future developments in neural networks will focus on integrating different generation methods and improving training datasets to enhance content quality.The practical significance of the research lies in its potential to reduce video game development costs while increasing the level of detail and realism in virtual worlds, opening new prospects for the gaming industry.
Authors
- Yuliia Yermolaieva
Institutions
- Interactive Research and Development (PK)
Publication Details
- Journal
- Věda a perspektivy
- Published
- 2025-05-02
- DOI
- https://doi.org/10.52058/2695-1592-2025-4(47)-239-246
- Primary Topic
- Advanced Research in Systems and Signal Processing
- Type
- article
- Field-Weighted Citation Impact
- 0.00