AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

OpenMask3D: Open-Vocabulary 3D Instance Segmentation

Ayça Takmaz, Elisabetta Fedele, Robert Sumner, Marc Pollefeys, Federico Tombari, Francis Engelmann

NeurIPS 2023 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary semantic segmentation · 29 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
add SAM-refined 2D masks for cropping (ablation)OpenMask3D · ScanNet200 valAP: 12.9 -> 14.1Improved
add multi-scale crops (3 levels) (ablation)OpenMask3D · ScanNet200 valAP: 12.9 -> 14.3Improved
combine SAM masks with multi-scale crops (ablation)OpenMask3D · ScanNet200 valAP: 12.9 -> 15.4Improved
increase top-k views from 1 to 5 (ablation)OpenMask3D · ScanNet200 valAP: 13.6 -> 15.4Improved
increase top-k views from 5 to 10 (ablation)OpenMask3D · ScanNet200 valAP: 15.4 -> 15.4No clear effect
use 3 crop levels instead of 1 (ablation)OpenMask3D · ReplicaAP: 11.3 -> 13.1Improved
use 5 crop levels instead of 3 (ablation)OpenMask3D · ReplicaAP: 13.1 -> 12.8Hurt
expansion ratio 0.05 or 0.2 instead of 0.1 (ablation)OpenMask3D · ReplicaAP: 13.1 -> 12.9 / 12.8Hurt
use ground truth oracle masks instead of predicted masksOpenMask3D · ScanNet200 valAP: 15.4 -> 29.1Improved
replace real RGB-D images with images rendered from the point cloudOpenMask3D · ReplicaAP: 13.1 -> 11.6Hurt
instance-mask CLIP features instead of OpenScene point features averaged over masksOpenMask3D · ScanNet200 valAP: 11.7 -> 15.4Improved
train mask module on ScanNet20 instead of ScanNet200OpenMask3D · ScanNet200 valAP: 15.4 -> 12.6Hurt

Later papers that compared with this method

7 later accepted papers report OpenMask3D as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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