AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Total Denoising: Unsupervised Learning of 3D Point Cloud Cleaning

Pedro Hermosilla, Tobias Ritschel, Timo Ropinski

ICCV 2019 · Point clouds › Shape and surface completion · approach: Point cloud upsampling and normals · 143 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 spatial proximity prior to unsupervised training (ablation)Total Denoising · Simple noise, 22M pointsChamfer error: .582 -> .547Improved
add color (appearance) to the prior sampling (ablation)Total Denoising · Simple noise, 22M pointsChamfer error: .547 -> .542Improved
add spatial prior (ablation)Total Denoising · Advanced noiseChamfer error: .393 -> .359Improved
add color to prior (ablation)Total Denoising · Advanced noiseChamfer error: .359 -> .356Improved
unsupervised Full instead of supervised trainingTotal Denoising · Simple noise, 22M pointsChamfer error: .545 -> .542Improved
unsupervised Full instead of supervised trainingTotal Denoising · Advanced noiseChamfer error: .356 vs .329Hurt
use unsupervised training with little data (0.5M points) instead of supervisedTotal Denoising · Simple noise, 0.5M pointsChamfer error: .574 -> .558Improved
increase training data from 0.5M to 22M points (ablation)Total Denoising Full · Simple noiseChamfer error: .558 -> .542Improved
use unsupervised Full instead of supervised at 1% noiseTotal Denoising · Simple noise 1.0%Chamfer error: .525 vs .515Hurt
use Gaussian kernel instead of Wendland or inverse multi-quadric for the prior (ablation)Total Denoising · Prior kernel ablationChamfer error: .548 vs .560 vs .577Improved
train on a single noise level instead of all levels (ablation)Total Denoising · Test 1.0% noiseChamfer error: .534 -> .551Hurt
apply our unsupervised prior to PointCleanNet trainingPCNet · PointCleanNet dataseterror: 1.36 -> 1.34No clear effect

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