AI Paper Atlas

Research map / Point clouds / Shape and surface completion

CAP-UDF: Learning Unsigned Distance Functions Progressively From Raw Point Clouds With Consistency-Aware Field Optimization

Junsheng Zhou, Baorui Ma, Shujuan Li, Yu-Shen Liu, Yi Fang, Zhizhong Han

TPAMI 2024 · Point clouds › Shape and surface completion · approach: Distance-function surface reconstruction · 79 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 consistency-aware Chamfer loss instead of Neural-Pull loss (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.2381 -> 0.1112Improved
keep abs projection instead of exponent output (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1218 -> 0.1112Improved
train second stage progressively end to end instead of from scratch (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1497 -> 0.1112Improved
add second stage of progressive surface approximation (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1218 -> 0.1112Improved
add third and fourth stages (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1112 -> 0.1107No clear effect
set low-confidence range to 1.1 sigma vs 0.9, 1.0, 1.2 (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1133, 0.1130, 0.1112, 0.1131Mixed
add mesh refinement at 256^3 (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1294 -> 0.1112Improved
add mesh refinement at 64^3 (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.4169 -> 0.1606Improved
increase grid resolution 128^3 to 256^3 (with refinement) (ablation)CAP-UDF · ShapeNet Cars subsetL2CD (x1e4): 0.1174 -> 0.1112; time 21.9 s -> 162.2 sImproved
stage-1 iterations 30k/40k/50k (ablation)CAP-UDF · ShapeNet Cars subsetChamfer-L2: 0.1158, 0.1112, 0.1137Mixed
stage-2 iterations 15k/20k/25k (ablation)CAP-UDF · ShapeNet Cars subsetChamfer-L2: 0.1125, 0.1112, 0.1133Mixed
estimate normals from fused query gradients (K=50) instead of point gradient (ablation)CAP-UDF · PCPNetnormal RMSE average: 7.61 -> 6.90Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

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