SpaCeFormer: Open-Vocabulary 3D Instance Segmentation Without 2D Images Goes Open Source

Accepted by ICML 2026, SpaCeFormer introduces open-vocabulary 3D instance segmentation directly from point clouds without 2D images. The project has fully open-sourced its code, weights, and interactive demo tools.

2026-07-08 ~ 2026-07-08 · 2 related posts

• SpaCeFormer: Interactive Open-Vocabulary 3D Instance Segmentation Without 2D Images — realChrisChoy · 2026-07-08
• SpaCeFormer Officially Open-Sourced: ICML 2026 Accepted with Full Release — realChrisChoy · 2026-07-08