AI Paper Atlas

Research map / Point clouds / Shape and surface completion

3D Shape Generation and Completion Through Point-Voxel Diffusion

Linqi Zhou, Yilun Du, Jiajun Wu

ICCV 2021 · Point clouds › Shape and surface completion · approach: Point cloud diffusion generation · 482 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 point-voxel representation instead of pure voxel diffusion (ablation)PVD · ShapeNet Airplane generation1-NN accuracy (CD): 99.75 -> 73.82Improved
use point-voxel representation instead of pure voxel diffusion (ablation)PVD · ShapeNet Chair generation1-NN accuracy (CD): 97.12 -> 56.26Improved
use point-voxel representation instead of pure voxel diffusion (ablation)PVD · ShapeNet Car generation1-NN accuracy (EMD): 96.83 -> 53.83Improved
use point-voxel CNN backbone instead of pure point diffusion with PointNet++ (ablation)PVD · ShapeNet generationvisual qualityImproved
use PVD instead of PointFlowPointFlow · ShapeNet Chair generation1-NN accuracy (CD): 62.84 -> 56.26Improved
use PVD instead of PointFlowPointFlow · ShapeNet Car generation1-NN accuracy (EMD): 56.25 -> 53.83Improved
use PVD instead of PointFlow for completionPointFlow · ShapeNet Chair completionCD x10^3 and EMD x10^2: CD 2.707 -> 3.211; EMD 3.649 -> 2.939Mixed
use PVD instead of PointFlow for completionPointFlow · ShapeNet Airplane completionCD x10^3 and EMD x10^2: CD 0.4030 -> 0.4415; EMD 1.180 -> 1.030Mixed
use PVD instead of DPF-Net for completionDPF-Net · ShapeNet Car completionCD x10^3 and EMD x10^2: CD 1.396 -> 1.774; EMD 2.318 -> 2.146Mixed
use PVD instead of cGAN for multi-modal completioncGAN · PartNet multi-modal completion (avg)TMD x10^2 (diversity): 1.89 -> 3.18Improved
use PVD instead of cGAN for multi-modal completioncGAN · PartNet multi-modal completion (avg)MMD x10^3 (quality): 1.77 -> 1.43Improved
use PVD instead of KNN-latent for multi-modal completionKNN-latent · PartNet multi-modal completion (avg)MMD x10^3 (quality): 1.57 -> 1.43No 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