AI Paper Atlas

Research map / Point clouds / Shape and surface completion

AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers

Xumin Yu, Yongming Rao, Ziyi Wang, Jiwen Lu, Jie Zhou

TPAMI 2023 · Point clouds › Shape and surface completion · approach: Structure-guided completion · 208 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 query generator between encoder and decoder (ablation)vanilla Transformer completion model · PCNCD-l1: 9.43 -> 9.09Improved
add DGCNN feature extractor for point proxies (ablation)Transformer with query generator · PCNCD-l1: 9.09 -> 8.69Improved
add geometric block to all Transformer blocks (ablation)Transformer with DGCNN · PCNCD-l1: 8.69 -> 8.44Improved
add geometric block to only the first block of encoder and decoder instead of all (ablation)PoinTr · PCNCD-l1: 8.44 -> 8.38Improved
add adaptive query generation (ablation)PoinTr · PCNCD-l1: 8.38 -> 7.06Improved
add dynamic query selection (ablation)PoinTr with adaptive queries · PCNCD-l1: 7.06 -> 6.92Improved
add auxiliary denoising task (ablation)AdaPoinTr · PCNCD-l1: 6.92 -> 6.53Improved
add adaptive queries, selection and denoising to PoinTrPoinTr · ShapeNet-55CD-l2 avg (x1000): 1.09 -> 0.81Improved
add adaptive queries, selection and denoising to PoinTrPoinTr · ShapeNet-34 (unseen 21 categories)CD-l2 avg (x1000): 2.05 -> 1.23Improved
add adaptive queries, selection and denoising to PoinTrPoinTr · Projected-ShapeNet-55CD-l1 (x1000): 10.68 -> 9.58Improved
add adaptive components to PoinTrPoinTr · MVP validation (2048 points)CD-l2 (x10000): 6.15 -> 4.71Improved
use PoinTr (depth only) geometry-enhanced Sketch-GE vs SketchSketch · NYUCAD semantic scene completionmIoU: 55.2 -> 56.1No clear effect

Later papers that compared with this method

16 later accepted papers report AdaPoinTr as a baseline in their results tables; 12 of them report a better number on most of the results they share with it.

8 more not listed.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 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