Keenan Crane

@keenancrane.bsky.social

Digital Geometer, Associate Professor of Computer Science & Robotics at Carnegie Mellon University. There are four lights. https://www.cs.cmu.edu/~kmcrane/

The wonderful Nicole Feng, defending her PhD thesis on generalized distances & winding numbers. So proud of her, and very excited about all the places she will go. nzfeng.github.io (Also that text outline shaded by the surface is some galaxy brain diagramming. 🤯🧠🌌)

BildBild

“Everyone knows” what an autoencoder is… but there's an important complementary picture missing from most introductory material. In short: we emphasize how autoencoders are implemented—but not always what they represent (and some of the implications of that representation).🧵

Bild

I can't* fathom why the top picture, and not the bottom picture, is the standard diagram for an autoencoder. The whole idea of an autoencoder is that you complete a round trip and seek cycle consistency—why lay out the network linearly?

Bild

I don't have a strong opinion about whether video models “understand the world.” But I do think the first bar should be checking whether you can recover consistent geometry from video—not whether it makes accurate predictions of physics.

Bild

Quick “teaser” for a fun #SIGGRAPH2025 project, led by Hossein Baktash, on optimizing a shape to have the desired rolling statistics. Basically we can turn arbitrary objects into fair dice, or make dice which capture the statistics of other objects—like several coin flips.

Making good on this promise—in the fastest turnaround time ever—my collaborator Etienne Corman has just posted MATLAB code for #RectangularSurfaceParameterization here: github.com/etcorman/Rec... (C++ version is still in the works…)

A teaser image for Rectangular Surface Parameterization, with large letters "RSP" meshed in blue, surrounded by a collection of yellow rectangular quad meshes generated by the RSP algorithm.
Keenan Crane@keenancrane.bsky.social · last yr.

Code and other information coming soon; for now you can read the paper here: www.cs.cmu.edu/~kmcrane/Pro... And find some supplemental information—including pseudocode—here: www.cs.cmu.edu/~kmcrane/Pro...

Meshes with 90° angles are super useful, providing asymptotically faster convergence for finite element simulation, and optimal shape approximation (when aligned with curvature). Amazingly, no past quad meshing method could guarantee 90° angles under refinement—until now. #RSP

Very happy to receive a SIGGRAPH 2025 Best Paper Award (Honorable Mention) for our RSP algorithm, which dices surfaces into near-perfect rectangles. Such meshes are useful for everything from retopology, to microfluidic simulation, to textile design, to architectural geometry.

A preview of three pages and a teaser image from the SIGGRAPH 2025 paper "Rectangular Surface Parameterization" by Etienne Corman & Keenan Crane, as well as a trophy with the SIGGRAPH logo embossed on it and the text "Best Paper Award (Honorable Mention)."

Shout out today to parents, grandparents, aunts, uncles, or anyone else who helps raise a kid. It didn’t really click until I was a dad: without parents, human civilization ends. Only with good parents does the future look bright. So, give a nod to anyone putting in the work.

Bild

Anyone interested in 3D printing some funky dice—like a "fair dragon" which can land in three different ways, or a single die equivalent to to 2D6? We've now put printable STL files online: www.cs.cmu.edu/~kmcrane/Pro... More info here: hbaktash.github.io

Bild
Keenan Crane@keenancrane.bsky.social · last yr.

Fun new paper at #SIGGRAPH2025: What if instead of two 6-sided dice, you could roll a single "funky-shaped" die that gives the same statistics (e.g, 7 is twice as likely as 4 or 10). Or make fair dice in any shape—e.g., dragons rather than cubes? That's exactly what we do! 1/n

Fun new paper at #SIGGRAPH2025: What if instead of two 6-sided dice, you could roll a single "funky-shaped" die that gives the same statistics (e.g, 7 is twice as likely as 4 or 10). Or make fair dice in any shape—e.g., dragons rather than cubes? That's exactly what we do! 1/n

Bild