iiinova — where your digital self keeps evolving

Yuezhi Yang @yyuezhi.bsky.social · 03/06/2026

Code is out! Feel free to try it here: github.com/nv-tlabs/lea... We’ll also be presenting our work as a CVPR oral at Bluebird Ballroom on Sunday, June 7, at 3:00 PM. Hope to see you there!"

github.com

GitHub - nv-tlabs/learning-convex-decomposition

Contribute to nv-tlabs/learning-convex-decomposition development by creating an account on GitHub.

000 0

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Check out the project page for more results and details! Many thanks to my amazing collaborators: Qixing Huang, Mikaela Uy, and @nmwsharp.bsky.social — and to the incredible NVIDIA Spatial Intelligence Lab for wonderful internship experience. 8/n

000 0

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

The resulting convex components can be seamlessly integrated into simulators such as MuJoCo and NVIDIA Newton for fast collision handling. 7/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Our hierarchical decomposition algorithm also supports granularity control. It allows users to interactively adjust the threshold to achieve the desired level of detail. 6/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Our model shows strong open-world generalizability across a diverse range of data in various modalities, including CAD models, 3D scans, and gaussian splats. 5/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Datasets of convex decompositions do not exist to learn from — that’s where our key idea comes in. We introduce an unsupervised contrastive loss derived from the definition of convexity, which pulls features of convex pairs of points together while pushing features of non-convex pairs apart. 4/n

110 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

We train a model that converts any shape into a feature field, where the distance between two points indicates whether they should belong to the same convex component. We then apply clustering on this field to obtain a hierarchical convex decomposition. 3/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Convex decomposition breaks a non-convex 3D shape into a set of convex components. It is widely used in physics engines as it accelerates geometric computation like collision detection. However, existing approaches either rely on expensive heuristic search or generalize poorly to openworld shape.2/n

110 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Excited to share our new work at CVPR 2026: Learning Convex Decomposition via Feature Fields. We introduce the first feedforward openworld model that generates high-quality convex decompositions for any 3D shapes in seconds, enabling faster simulation. Project: research.nvidia.com/labs/sil/pro...

1136 13

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

000 0

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

The resulting convex components can be seamlessly integrated into simulators such as MuJoCo and NVIDIA Newton for fast collision handling. 6/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Our model shows strong open-world generalizability across a diverse range of data in various modalities, including CAD models, 3D scans, and gaussian splats. 5/n

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

100 1

Yuezhi Yang @yyuezhi.bsky.social · 10/03/2026

Convex decomposition breaks a non-convex 3D shape into a union of convex components.It is widely used in physics engines as it accelerates geometric computations like collision detection. However, existing approach either rely on expensive heuristic search or generalize poorly to open-world shapes.

110 1