AI Paper Atlas

Research map / Point clouds / Point cloud classification

SampleNet: Differentiable Point Cloud Sampling

Itai Lang, Asaf Manor, Shai Avidan

CVPR 2020 · Point clouds › Point cloud classification · approach: Efficient point sampling architectures · 204 citations (OpenAlex)

Check 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
train a task-specific sampler with differentiable soft projection, then sample from the input at inferencePointNet classification · ModelNet40 test, sampling ratio 32classification accuracy: 80.1%Improved
use SampleNet instead of FPS for point-cloud registrationPCRNet · ModelNet40 Car test, sampling ratio 32mean rotation error: $13.46\degree$ -> $5.94\degree$Improved
use SampleNet instead of S-NET for point-cloud reconstructionPCAE · ShapeNet Core55 test, sampling ratio 32normalized reconstruction error: 1.57 -> 1.33Improved
use a decaying temperature profile rather than a constant temperature (ablation)SampleNet-Progressive classification · ModelNet40 test, sampling ratio 32classification accuracy: 17.1 -> 74.6Improved
use a linear rectified or exponential temperature schedule instead of the learned schedule (ablation)SampleNet-Progressive classification · ModelNet40 testclassification accuracyNo clear effect
add cross-entropy loss to force projection weights toward the nearest neighbor (ablation)SampleNet classification · ModelNet40 test, sampling ratio 32classification accuracy: 80.1 -> 74.4Hurt
add cross-entropy loss to force projection weights toward the nearest neighbor (ablation)SampleNet classification · ModelNet40 test, sampling ratio 2classification accuracy: 88.4 -> 88.2Hurt
add entropy loss to force projection weights toward a one-hot distribution (ablation)SampleNet classification · ModelNet40 testclassification accuracyHurt
use a projection neighborhood size other than k = 7 (ablation)SampleNet-Progressive classification · ModelNet40 testclassification accuracyHurt
use FPS instead of SampleNet for airplane reconstructionPCAE · ShapeNet Core55 airplane, sample size 64normalized reconstruction error: 1.39 -> 1.31Improved
use SampleNet for registration rather than FPS as sampling becomes sparserPCRNet · ModelNet40 Car test, sampling ratio 32sampling consistency: 11.78 -> 6.85Improved

Later papers that compared with this method

4 later accepted papers report SampleNet as a baseline in their results tables; 4 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