DTU Benchmarks

[Korace0v0]

Hi, thanks for the impressive work!

I’m really intrigued by the approach of extracting meshes directly from geometry fields. Do you have any plans to release the training and evaluation scripts for the DTU dataset?

Looking forward to your response!

[sel8888]

Hello, great work,
I discovered that a single Laplace-Beltrami eigenfunction
can speed up and improve ANY spectral shape correspondence method.

+42% ZoomOut. +52% Functional Maps. 99 FAUST pairs. 0.8s. No training.

Looking for collaborators to take this further

github.com/sel8888/harmonic-shape-transform-2026-koncept
doi.org/10.5281/zenodo.20059602
0009-0003-9680-3333

hst_faust_full_benchmark.csv

[zhangxujinsh]

@diego1401
Looking forward to your response for DTU dataset!
Thanks for your great work!