Workshop on 3D Geometry Generation for Scientific Computing

The Workshop on 3D Geometry Generation for Scientific Computing aims to bring together the AI and scientific communities to discuss the state of the art in 3D geometry generation and how it can be applied to open problems in science. We welcome methods for 3D geometry generation as well as applications of 3D geometries in a scientific context. The workshop is non-archival, so while we encourage submission of novel work, previously published work is acceptable. Selected abstracts may be included (with prior consent) in a workshop summary paper in the WACV 2024 proceedings.

Submission Deadline: Nov 2nd, 2023 Nov 9th, 2023

Notifications: Nov 9th, 2023  Nov 16th, 2023

Submit to Open Review: https://openreview.net/group?id=thecvf.com/WACV/2024/Workshop/3D4Science

We provide three tracks for submission of work:

(A) Methods in Geometry Generation

This track is for methods of geometry generation that are application agnostic. While methods do not have to be tied to any particular scientific problem, papers must illustrate how the method can be used to create a geometry of scientific interest (i.e. an outdoor scene). 

Example topics include (but are not limited to):

Reconstruction of large-scale outdoor scenes

Uncertainty quantification in geometry

Reconstruction from sparse data; Scientifically informed shape completion

(B) Applications of Geometric Reconstruction

This track is for novel applications of 3D geometric reconstruction in the scientific domain. 

Example topics include (but are not limited to):

Open datasets of real-world geometries (particularly forests, oceans, glaciers, etc.)

Simulation with neural geometries

Physics informed methods of geometry estimation

Tools and libraries for building geometries for scientific workflows

(C) Geometric Challenge Problems (Extended Abstract)

This track is an opportunity to motivate work in 3D geometry reconstruction by identifying a geometry (or class of geometries) of high scientific value that is not currently available or not available at high enough fidelity. 

Papers must clearly state (1) what high-value scientific questions could be answered if this geometry were available (2) what technical problems make it so that current state-of-the-art methods are not able to reconstruct this geometry to high enough accuracy. 

Style & Author Instructions