August 10th, 2024

Ray Tracing Harmonic Functions

A paper titled "Ray Tracing Harmonic Functions" by Mark Gillespie and colleagues received an Honorable Mention at SIGGRAPH 2024 for introducing a novel sphere tracing algorithm for efficient surface visualization.

Read original articleLink Icon
Ray Tracing Harmonic Functions

A recent paper titled "Ray Tracing Harmonic Functions," authored by Mark Gillespie and colleagues, received an Honorable Mention for Best Paper at SIGGRAPH 2024. The research introduces a novel sphere tracing algorithm specifically designed for harmonic functions, which are significant in geometric and visual computing as solutions to interpolation problems and the Laplace equation. This new algorithm allows for efficient visualization of surfaces represented by harmonic functions, leveraging Harnack's inequality to establish a conservative lower bound on the distance to the level set. This approach enables larger step sizes compared to traditional ray marching methods. The algorithm can effectively visualize smooth surfaces reconstructed from point clouds and polygonal data without the need for complex linear solves or mesh extraction. Additionally, it can render nonplanar polygons and visualize mathematical objects such as knots and Riemann surfaces. The paper includes practical implementations in GLSL shaders, demonstrating the algorithm's capabilities in various applications, including point cloud reconstruction and the visualization of hyperspherical harmonics. The research was supported by multiple grants and collaborations, highlighting its significance in advancing techniques in computer graphics.

- The paper presents a new sphere tracing algorithm for harmonic functions.

- It allows for efficient visualization of surfaces without complex mesh extraction.

- The algorithm utilizes Harnack's inequality for improved performance over traditional methods.

- Practical implementations are provided in GLSL shaders for various applications.

- The research received an Honorable Mention at SIGGRAPH 2024.

Link Icon 2 comments
By @pixelpoet - 6 months
Keenan Crane back once again with the ~~renegade master~~ shithot diagrams! For those not familiar, check his publications with ridiculously beautiful diagrams, going way back (I remember when he was maybe best known for a Quaternion Julia tracer, often used to benchmark GPU FLOPs): https://www.cs.cmu.edu/~kmcrane/

Also fun for topological nightmare enjoyers is his "Yeah Right" model on https://www.cs.cmu.edu/~kmcrane/Projects/ModelRepository/ (img link: https://www.cs.cmu.edu/~kmcrane/Projects/ModelRepository/yea...)

By @nimish - 6 months
Cool as hell. Keenan Crane has excellent YouTube series on discrete differential geometry and computer graphics.

Hopefully this will lead to nicer physics graphics since harmonic functions are everywhere.