SpaCeFormer: Fast Proposal-Free Open-Vocabulary 3D Instance Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Choy, Chris, Lee, Junha, Park, Chunghyun, Cho, Minsu, Kautz, Jan
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910271225724928
author Choy, Chris
Lee, Junha
Park, Chunghyun
Cho, Minsu
Kautz, Jan
author_facet Choy, Chris
Lee, Junha
Park, Chunghyun
Cho, Minsu
Kautz, Jan
contents Open-vocabulary 3D instance segmentation is a core capability for robotics and AR/VR, but prior methods trade one bottleneck for another: multi-stage 2D+3D pipelines aggregate foundation-model outputs at hundreds of seconds per scene, while pseudo-labeled end-to-end approaches rely on fragmented masks and external region proposals. We present SpaCeFormer, a proposal-free space-curve transformer that runs in 0.12--0.30 seconds per scene across standard benchmarks, 2--3 orders of magnitude faster than multi-stage 2D+3D pipelines. We pair it with SpaCeFormer-3M, the largest open-vocabulary 3D instance segmentation dataset (3.0M multi-view-consistent captions over 604K instances from 7.4K scenes) built through multi-view mask clustering and multi-view VLM captioning; it reaches 21$\times$ higher mask recall than prior single-view pipelines (54.3% vs 2.5% at IoU$>$0.5). SpaCeFormer combines spatial window attention with Morton-curve serialization for spatially coherent features, and uses a RoPE-enhanced decoder to predict instance masks directly from learned queries without external proposals. On ScanNet200 we achieve 11.1 zero-shot mAP, a 2.8$\times$ improvement over the prior best proposal-free method; on ScanNet++ and Replica, we reach 22.9 and 24.1 mAP, surpassing all prior methods including those using multi-view 2D inputs.
format Preprint
id arxiv_https___arxiv_org_abs_2604_20395
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SpaCeFormer: Fast Proposal-Free Open-Vocabulary 3D Instance Segmentation
Choy, Chris
Lee, Junha
Park, Chunghyun
Cho, Minsu
Kautz, Jan
Computer Vision and Pattern Recognition
Robotics
Open-vocabulary 3D instance segmentation is a core capability for robotics and AR/VR, but prior methods trade one bottleneck for another: multi-stage 2D+3D pipelines aggregate foundation-model outputs at hundreds of seconds per scene, while pseudo-labeled end-to-end approaches rely on fragmented masks and external region proposals. We present SpaCeFormer, a proposal-free space-curve transformer that runs in 0.12--0.30 seconds per scene across standard benchmarks, 2--3 orders of magnitude faster than multi-stage 2D+3D pipelines. We pair it with SpaCeFormer-3M, the largest open-vocabulary 3D instance segmentation dataset (3.0M multi-view-consistent captions over 604K instances from 7.4K scenes) built through multi-view mask clustering and multi-view VLM captioning; it reaches 21$\times$ higher mask recall than prior single-view pipelines (54.3% vs 2.5% at IoU$>$0.5). SpaCeFormer combines spatial window attention with Morton-curve serialization for spatially coherent features, and uses a RoPE-enhanced decoder to predict instance masks directly from learned queries without external proposals. On ScanNet200 we achieve 11.1 zero-shot mAP, a 2.8$\times$ improvement over the prior best proposal-free method; on ScanNet++ and Replica, we reach 22.9 and 24.1 mAP, surpassing all prior methods including those using multi-view 2D inputs.
title SpaCeFormer: Fast Proposal-Free Open-Vocabulary 3D Instance Segmentation
topic Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2604.20395