RENDERING OPTIMIZATION METHOD FOR DYNAMIC SCENES USING AN ADAPTIVE HIERARCHY OF BOUNDING VOLUMES
Andrei P. Titov, Dar’ya N. Titova · RSUH/RGGU Bulletin Series Information Science Information Security Mathematics · 2025
The article considers a new approach to the algorithm for constructing and optimizing hierarchies of bounding volumes (BVH). Such a method is based on the analysis of the distribution of the objects boundaries in space. The relevance of the chosen topic is due to the rapid development of visualization and modeling technologies, especially in the field of computer graphics, animation and computer vision. Effective geometry management in three-dimensional space is critical to improve rendering performance and work with large amounts of data. The novelty of the proposed approach lies in the combination of methods for dynamic distribution and adaptive modification of the BVH structure depending on the characteristics of the input data. Thus it enables to significantly improve the processing speed of scenes containing complex and highly detailed objects. Unlike traditional methods that rely on static algorithms, the considered method adapts in real time to changes in the geometry and distribution of objects. The main advantages of the method are its high efficiency and versatility. Adaptive algorithms improve the quality and speed of ray intersection with limiting volumes, which significantly reduces rendering time. Due to the consideration of changes in the distribution of objects, the method shows high resilience to scenarios with dynamic or interactive scenes. Hierarchies of bounding volumes constructed using that approach can be effectively used from video games to scientific modeling. The method demonstrates its applicability to the creation of both static and adaptive hierarchy maps, which makes it suitable for multi-user and virtual environments. The boundary distribution-based BVH algorithm represents a significant step forward in terms of performance and functionality, opening up new horizons for future research and practical applications in the field of visualization and data processing.