atlas

Research map / 3D vision / 3D generation

3D-SIS: 3D Semantic Instance Segmentation of RGB-D Scans

CVPR 2019 · 3D vision › 3D generation · approach: Semantic scene completion

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 multi-view RGB features to geometry features, using 5 views (ablation)3D-SIS · ScanNetV2mAP@0.5: 16.0 -> 22.5Helped
increase the number of RGB views used with geometry from 1 to 5 (ablation)3D-SIS · ScanNetV23D detection mAP@0.25: 35.1 -> 40.2Helped
increase the number of RGB views used with geometry from 1 to 5 (ablation)3D-SIS · SUNCG3D detection mAP@0.25: 30.9 -> 32.2Helped
use 5 color views rather than geometry only (ablation)3D-SIS · SUNCGinstance segmentation mAP@0.25: 26.8 -> 30.6Helped
use 5 color views rather than geometry only (ablation)3D-SIS · ScanNetV2instance segmentation mAP@0.25: 28.3 -> 35.7Helped
replace the two-level anchor design with one anchor layer (ablation)3D-SIS · ScanNetV2mAP@0.5: 22.5 -> 12.2Hurt
train on color-only input with 5 views instead of geometry-only input (ablation)3D-SIS · ScanNetV2mAP@0.5: 16.0 -> 17.4Helped

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