580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can parametric and implicit polynomial surface representations be optimally converted and utilized for geometric modeling tasks?
This research area addresses the challenges and methods for representing surfaces with polynomial bases, focusing on parametric and implicit forms. It is crucial for efficient data exchange, surface modeling, mesh generation, and rendering in computer-aided design (CAD) and graphics. The theme encompasses approximate conversion techniques between different polynomial bases and analyzes the computational benefits and trade-offs of each representation for various geometric operations.
Key finding: The paper presents methods to approximate conversion between polynomial surface representations with differing bases, polynomial degrees, and mesh sizes, enabling effective data exchange between CAD systems. It emphasizes the... Read more
Key finding: This work compares implicit and parametric polynomial surface representations from an algebraic geometry perspective, highlighting their advantages in rendering and mesh generation. It asserts parametric forms, viewed as... Read more
2. What numerical methods enable bending-invariant and isometry-invariant surface representations for shape matching and visualization?
Research under this theme investigates surface representations invariant under isometric deformations and bending, focusing on exploiting intrinsic geometric properties such as geodesic distances. These representations facilitate robust surface matching, classification, and visualization that is resilient to deformations preserving intrinsic metric, which is vital for applications from medical imaging to computer graphics.
Key finding: The authors introduce a method constructing bending-invariant canonical forms by embedding intrinsic geodesic distances of isometric surfaces into finite-dimensional Euclidean spaces via efficient fast marching on... Read more
Key finding: This work proposes a generalized model (SMURF) to access up to second-order surface properties (positions, normals, curvature) independent of surface origin, allowing visualization algorithms to process diverse surface types... Read more
3. How can high-performance, interactive visualization of complex and dynamic surfaces be achieved using implicit representations and GPU computing?
This theme explores novel computational methods and algorithms focused on fast, interactive surface visualization and deformation, especially for implicit surfaces and dynamic data such as molecular structures or volumetric images. It emphasizes GPU-based acceleration, level-set methods, and real-time surface generation to facilitate exploratory analysis and scientific visualization.
Key finding: The paper presents a grid-based, GPU-accelerated algorithm for on-the-fly solvent-excluded surface (SES) generation suitable for large, time-varying molecular data. By decomposing SES computation into two phases and... Read more
Key finding: This study introduces a GPU-accelerated PDE solver for level-set methods applicable to deforming implicit surfaces. Achieving approximately 15x speedup over optimized CPU solutions, it combines interactive level-set... Read more
Key finding: This survey classifies and compares various polygonization techniques for fast visualization of implicit surfaces, highlighting trade-offs between accuracy, speed, and mesh quality. It identifies polygonization as the main... Read more