AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Point Diffusion Mamba: Unified Diffusion-State-Space Modeling for Single-View 3D Reconstruction under Data Scarcity

Wei Zhou, Xinzhe Shi, Xingxing Hao, Xing Hao, Kang Li, Jinye Peng, Ying He

ECCV 2026 · Point clouds › Shape and surface completion · approach: Point cloud diffusion generation · 0 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 Local Geometric Aggregation (ablation)PDM · Pix3D ChairF1@0.01: 0.478 -> 0.499Improved
add Hierarchical Feature Integration Network (ablation)PDM · Pix3D ChairF1@0.01: 0.435 -> 0.499Improved
use Mamba blocks instead of self-attention (ablation)PDM · Pix3D ChairF1@0.01: 0.481 -> 0.499Improved
use dual-directional SSM instead of one SSM (ablation)PDM · Pix3D ChairF1@0.01: 0.474 -> 0.499Improved
use Dynamic Weighted Sampling instead of direct sampling (ablation)PDM · Pix3D ChairChamfer Distance: 85.27 -> 79.28Improved
use DWS instead of BDM sampling (ablation)PDM · Pix3D ChairF1@0.01: 0.492 -> 0.499Improved
input patch length 32 to 128 (ablation)PDM · Pix3D ChairF1@0.01: 0.485 -> 0.499Improved
lengthen patch sequence beyond 128 (to 384) (ablation)PDM · Pix3D ChairF1@0.01: 0.499 -> 0.487Hurt
scale model from PDM-B to PDM-L (ablation)PDM · Pix3D ChairF1@0.01: 0.499 -> 0.492No clear effect
use PDM instead of MESC-3D at 10% training dataPDM · ShapeNet-R2N2 Chair (10% data)Chamfer Distance: 101.51 -> 82.41Improved
use PDM instead of CCD-3DR at 100% dataPDM · ShapeNet-R2N2 Chair (100% data)Chamfer Distance: 58.47 -> 62.14Hurt
use PDM instead of MESC-3D at 50% dataPDM · ShapeNet-R2N2 Car (50% data)Chamfer Distance: 44.48 -> 55.53Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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