AI Paper Atlas

Research map / Point clouds / Point cloud classification

Fully-Convolutional Point Networks for Large-Scale Point Clouds

Dario Rethage, Johanna Wald, Jurgen Sturm, Nassir Navab, Federico Tombari

ECCV 2018 · Point clouds › Point cloud classification · approach: Local feature aggregation networks · 216 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 FCPN instead of voxel-based ScanNet networksemantic voxel labeling · ScanNet semantic voxel labelingweighted accuracy: 73.0% -> 82.6%Improved
use FCPN instead of voxel-based ScanNet networksemantic voxel labeling · ScanNet semantic voxel labelingunweighted accuracy: 50.8% -> 54.2%Improved
use FCPN instead of PointNet++ MSG+DPsemantic voxel labeling · ScanNet semantic voxel labelingweighted accuracy: 84.5% -> 82.6%Hurt
use FCPN instead of PointNet++ MSG+DPsemantic voxel labeling · ScanNet semantic voxel labelingunweighted accuracy: 60.2% -> 54.2%Hurt
use FCPN instead of ScanNet baseline for chairssemantic voxel labeling · ScanNet semantic voxel labeling - Chairclass accuracy: 69.3% -> 81.6%Improved
use FCPN instead of ScanNet baseline for deskssemantic voxel labeling · ScanNet semantic voxel labeling - Deskclass accuracy: 36.8% -> 58.5%Improved
use FCPN instead of ScanNet baseline for doorssemantic voxel labeling · ScanNet semantic voxel labeling - Doorclass accuracy: 48.9% -> 16.6%Hurt
use FCPN instead of PointNet++ on part segmentationpart segmentation · ShapeNet part segmentationmean IoU: 85.1 -> 84.0Hurt
use FCPN instead of PointNet on part segmentationpart segmentation · ShapeNet part segmentationmean IoU: 83.7 -> 84.0Improved
use FCPN instead of PointNet++ for carspart segmentation · ShapeNet part segmentation - carIoU: 77.3 -> 88.3Improved
use FCPN instead of PointNet++ for tablespart segmentation · ShapeNet part segmentation - tableIoU: 82.6 -> 73.4Hurt
scale input from 15k points (16 m2) to 150k points (80 m2)FCPN inference · Titan Xp inferencememory: 6481 MB -> 9033 MBMixed

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