AI Paper Atlas

Research map / Point clouds / Point cloud classification

KPConv: Flexible and Deformable Convolution for Point Clouds

Hugues Thomas, Charles R. Qi, Jean-Emmanuel Deschaud, Beatriz Marcotegui, Francois Goulette, Leonidas J. Guibas

ICCV 2019 · Point clouds › Point cloud classification · approach: Local feature aggregation networks · 3,574 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 deformable kernels instead of rigidKPConv · ModelNet40overall accuracy: 92.9 -> 92.7Hurt
use deformable kernels instead of rigidKPConv · ShapeNetPartinstance mIoU: 86.2 -> 86.4Improved
use deformable kernels instead of rigidKPConv · S3DIS Area-5mIoU: 65.4 -> 67.1Improved
use deformable kernels instead of rigidKPConv · Paris-Lille-3DmIoU: 72.3 -> 75.9Improved
use deformable kernels instead of rigidKPConv · Scannet (test)mIoU: 68.6 -> 68.4Hurt
use deformable kernels instead of rigidKPConv · Semantic3D reduced-8mIoU: 74.6 -> 73.1Hurt
use deformable kernels instead of rigidKPConv · S3DIS (k-fold)mIoU: 69.6 -> 70.6Improved
reduce kernel points to 4 for deformable vs rigid (ablation)KPConv · Scannet validationmIoU: deformable loses 1.5% vs rigid loses 3.5%Improved
use deformable instead of rigid with 15 kernel points (ablation)KPConv · Scannet validationmIoUImproved
replace nearest upsampling by KPConv in decoder (ablation)KP-FCNN · segmentationmIoUNo clear effect
add fitting and repulsive regularization loss to deformable kernels (ablation)deformable KPConv · segmentationqualitativeImproved
use KPConv instead of other point convolutions on ScannetKP-FCNN · ScannetmIoU: exceed previous scores by 19 mIoU pointsImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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