AI Paper Atlas

Research map / 3D vision / 3D generation

Attention-DP3: Spatially Object-aware 3D Diffusion Policy via Geometry-aligned Attentional Conditioning

Changbo Yan, Zhongbo Zhang, Zaibin Zhang, Yifan Wang, Lijun Wang, Huchuan Lu

ECCV 2026 · 3D vision › 3D generation · approach: Score distillation for text-to-3D · 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 intra-target saliency and backgroundness to targetness-only (ablation)Attention-DP3 (T only) · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.5425 -> 0.72Improved
add targetness to intra-target saliency plus backgroundness (ablation)Attention-DP3 (B+I) · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.595 -> 0.72Improved
add intra-target saliency to targetness plus backgroundness (ablation)Attention-DP3 (T+B) · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.5625 -> 0.72Improved
add backgroundness to targetness plus saliency (ablation)Attention-DP3 (T+I) · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.5175 -> 0.72Improved
encode each field with its own AttnEncoder instead of a single encoder (ablation)Attention-DP3 · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.5225 -> 0.72Improved
use late fusion instead of early fusion (ablation)Attention-DP3 · Door/Stick-Pull/Pick-Place/Push-Wall avgsuccess rate: 0.255 -> 0.72Improved
add Tri-Field Attentional Conditioning to DP3DP3 · MetaWorld overallsuccess rate: 0.669 -> 0.726Improved
add Tri-Field Attentional Conditioning to DP3DP3 · Adroit avgsuccess rate: 0.68 -> 0.78Improved
add Tri-Field Attentional Conditioning to DP3DP3 · DexArt avgsuccess rate: 0.52 -> 0.56Improved
add Tri-Field Attentional Conditioning to DP3DP3 · SO101 real world avgsuccess rate: 0.52 -> 0.73Improved
add Tri-Field Attentional Conditioning under extra real-world distractorsDP3 · SO101 +Extra distractorssuccess rate: 0.13 -> 0.45Improved
replace predicted masks with ground-truth masks (ablation)Attention-DP3 · Mask perturbation avgsuccess rate: 0.524 -> 0.542No 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