Virtualización de nubes de puntos de gran tamaño
Thesis overview
Author: José-Antonio Collado-Araque
Supervision: Juan-Roberto Jiménez-Pérez, Alfonso López-Ruiz
Degree: Master's Degree in Computer Science
Institution: University of Jaén, Department of Computer Science
Defense: July 2024
Final grade: 10/10 (with honors)
Publications
- Virtualized Point Cloud Rendering
IEEE Transactions on Visualization and Computer Graphics, 2025 — DOI
Abstract
Visualizing large point clouds is challenging due to issues such as excessive memory requirements, inefficient point rasterization, and gaps in visual representations. To address these limitations, an approach based on meshlets—small geometry units—is proposed, enabling more efficient memory management and rendering performance optimizations. A methodology is designed, comprising a GPU-driven level of detail system, a virtual memory system responsible for loading and unloading points on demand, more efficient software rasterization than that provided by graphic APIs, and a gap-filling algorithm aimed at improving visualization by hiding some of these gaps. The proposed methodology can visualize clouds of several billion points, maintaining an average of 150 FPS without apparent quality loss.
