AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Noise2Score3D: Tweedie's Approach for Unsupervised Point Cloud Denoising

Xiangbin Wei, Yuanfeng Wang, Ao Xu, Lingyu Zhu, Dongyong Sun, Keren Li, Yang Li, Qi Qin

ICCV 2025 · Point clouds › Shape and surface completion · approach: Point cloud upsampling and normals · 6 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 AR-DAE loss instead of DAE loss (ablation)Noise2Score3D · ModelNet-40 50k, 1% noiseChamfer distance x1e4: 5.153 -> 1.881Improved
use AR-DAE loss instead of DAE loss (ablation)Noise2Score3D · ModelNet-40 50k, 3% noiseChamfer distance x1e4: 34.999 -> 8.917Improved
use AR-DAE loss instead of DAE loss (ablation)Noise2Score3D · ModelNet-40 50k, 2% noisepoint-to-mesh x1e4: 16.092 -> 4.002Improved
anneal sigma_t during training instead of fixed 0.01 (ablation)Noise2Score3D · ModelNet-40, 1% noiseChamfer distance x1e4: 2.014 -> 1.881Improved
anneal sigma_t during training instead of fixed 0.03 (ablation)Noise2Score3D · ModelNet-40, 3% noiseChamfer distance x1e4: 8.511 -> 8.917Hurt
anneal sigma_t during training instead of fixed 0.01 (ablation)Noise2Score3D · PU-Net, 1% noiseChamfer distance x1e4: 1.254 -> 0.833Improved
anneal sigma_t during training instead of fixed 0.01 (ablation)Noise2Score3D · PU-Net, 3% noiseChamfer distance x1e4: 2.395 -> 2.434Hurt
use TV_PC-estimated sigma instead of true sigma (ablation)Noise2Score3D · PU-Net 10K, 2% noiseChamfer distance x1e4: 3.868 -> 4.190Hurt
use TV_PC-estimated sigma instead of true sigma (ablation)Noise2Score3D · PU-Net 50K, 1% noiseChamfer distance x1e4: 0.991 -> 0.833Improved
use KPConv backbone instead of PointNet++ (ablation)Noise2Score3D · PU-Net 10K, 3% noiseChamfer distance x1e4: 13.382 -> 5.583Improved
use KPConv backbone instead of original backbone (ablation)DMR-U · PU-Net 10K, 1% noiseChamfer distance x1e4: 5.313 -> 33.89Hurt
use KPConv backbone instead of original backbone (ablation)Score-U · PU-Net 10K, 1% noiseChamfer distance x1e4: 3.107 -> 3.494Hurt

Later papers that compared with this method

1 later accepted paper reports Noise2Score3D as a baseline in their results tables; 1 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