Octree Structure
Octrees are hierarchical tree data structures that efficiently represent 3D data by recursively subdividing space into eight octants. Current research focuses on leveraging octrees for various applications, including 3D shape generation and reconstruction using diffusion models and neural networks, efficient scene representation for real-time rendering, and point cloud processing with state-space models and transformers. This adaptable structure offers significant advantages in memory efficiency, computational speed, and the ability to handle varying levels of detail, impacting fields like robotics, computer graphics, and cultural heritage preservation through improved 3D data management and processing.
Papers
November 29, 2024
November 25, 2024
August 27, 2024
April 27, 2024
April 5, 2024
March 26, 2024
March 21, 2024
March 11, 2024
December 6, 2023
September 15, 2023
March 22, 2023
December 12, 2022