AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Revisiting Point Cloud Completion: Are We Ready For The Real-World?

Stuti Pathak, Prashant Kumar, Dheeraj Baiju, Nicholus Mboga, Gunther Steenackers, Rudi Penne

ICCV 2025 · Point clouds › Shape and surface completion · approach: Point cloud completion networks · 7 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 0-dim persistent homology topological loss on decoder seeds (ablation)ODGNet · RealPC completionCD-L1 (x1e-3): 119 -> 103Improved
add 0-dim persistent homology topological loss on decoder seeds (ablation)ODGNet · RealPC completionCD-L2 (x1e-3): 111 -> 80Improved
replace PH computation with Homology Sampler (BOSH) proxy backbonesODGNet / TopoODG · RealPC completionCD-L2 (x1e-3): 111 -> 5.4Improved
replace PH computation with Homology Sampler (BOSH) proxy backbonesSnowflakeNet · RealPC completionCD-L1 (x1e-3): 60 -> 69No clear effect
add 0-dim PH priors to DGCNN-based generative model (ablation)DGCNN-based model · KITTI scene completionChamfer distance (x1e-3): 1.18 -> 1.04Improved
add 0-dim PH priors to DGCNN-based generative model (ablation)DGCNN-based model · KITTI scene reconstructionChamfer distance (x1e-3): 1.50 -> 1.18Improved
evaluate completion on real-world industrial point clouds instead of PCNODGNet · completion mean CD-L1 (x1e-3)CD-L1 (x1e-3): 6 -> 119Hurt
evaluate completion on real-world industrial point clouds instead of PCNAdaPoinTr · completion mean CD-L1 (x1e-3)CD-L1 (x1e-3): 7 -> 69Hurt
evaluate simplification on RealPC instead of ShapeNetWLOP simplification · non-neural simplificationChamfer distance (x1e-3): 0.522 -> 8.817Hurt
evaluate alpha-shape surface reconstruction on RealPC instead of ShapeNetalpha shapes · non-neural surface reconstructionHausdorff distance (x1e-3): 68.198 -> 494.591Hurt
evaluate reconstruction on RealPC instead of ShapeNetFoldingNet · reconstructionChamfer distance (x1e-3): 5 -> 27Hurt
evaluate diffusion shape generation on RealPC instead of ShapeNetLION · shape generation 1-NNA (CD)1-NNA: 67 -> 90Hurt

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