AI Paper Atlas

Research map / Point clouds / Segmentation and detection

Part2Object: Hierarchical Unsupervised 3D Instance Segmentation

Cheng Shi, Yulin Zhang, Bin Yang, Jiajin Tang, Yuexin Ma, Sibei Yang

ECCV 2024 · Point clouds › Segmentation and detection · approach: Scene completion and instance segmentation · 11 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use hierarchical multi-layer clustering instead of single-layer clustering (ablation)Part2Object · ScanNet valmAP@50: 13.8 -> 26.8Improved
add 3D objectness guidance (stop criterion) (ablation)Part2Object · ScanNet valmAP@50: 20.7 -> 26.8Improved
use similarity-weighted cluster features instead of plain average (ablation)Part2Object · ScanNet valmAP: 10.6 -> 12.6Improved
add object-part queries (Hi-Mask3D) over Mask3D (ablation)Mask3D on Part2Object pseudo-labels · ScanNet valmAP@50: 31.0 -> 36.0Improved
add rounds of self-training (ablation)Hi-Mask3D · ScanNet valmAP: 15.3 -> 16.9Improved
vary merge ratio K (0.4 to 0.7) (ablation)Part2Object · ScanNet valmAP@50: K=0.6 gives 26.8, K=0.7 gives 21.1, K=0.4 gives 22.6Mixed
use Part2Object pseudo-label pre-training instead of unsupervised baselines Unscene3DHi-Mask3D · ScanNet val (1% labels)mAP@50: 28.4 -> 44.1Improved
train on unsupervised Part2Object pseudo-labels instead of fully supervised ScanNet labelsHi-Mask3D · ScanNet200 zero-shotmAP@50: 24.2 -> 34.9Improved
pre-train on Part2Object pseudo-labels before data-efficient fine-tuningHi-Mask3D · ScanNet200 (1% data)mAP@50: 1.3 -> 35.4Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy