AI Paper Atlas

Research map / 3D vision / 3D generation

3D Instance Segmentation via Multi-Task Metric Learning

Jean Lahoud, Bernard Ghanem, Marc Pollefeys, Martin R. Oswald

ICCV 2019 · 3D vision › 3D generation · approach: Semantic scene completion · 161 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 directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · ScanNet v2 valAP50: 0.333 -> 0.402Improved
add directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · ScanNet v2 valAP: 0.171 -> 0.203Improved
add directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · ScanNet v2 valAP25: 0.492 -> 0.554Improved
add directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · Synthetic toy dataset, Obj1AP50: 97.3 -> 98.0Improved
add directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · Synthetic toy dataset, Obj2AP50: 92.7 -> 93.5Improved
add directional loss (multi-task) to feature embedding (ablation)3D instance segmentation network · Synthetic toy dataset, Obj5AP50: 95.2 -> 95.3No clear effect
add directional loss on the counter class (ablation)3D instance segmentation network · ScanNet v2 val, counterAP50: 0.004 -> 0.008No clear effect
replace segmentation-as-instance with learned feature embedding (FE only)3D instance segmentation · ScanNet v2 valAP50: 0.145 -> 0.333Improved
replace connected components with learned feature embedding (FE only)3D instance segmentation · ScanNet v2 valAP50: 0.230 -> 0.333Improved
apply connected components to semantic segmentation labelssemantic segmentation as instances · ScanNet v2 valAP50: 0.145 -> 0.230Improved
learned embedding vs connected components on ground-truth labelstoy data · Synthetic toy dataset, Obj5AP50: 79.9 -> 95.3Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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