AI Paper Atlas

Research map / 3D vision / Human pose estimation

TransGrasp: Grasp Pose Estimation of a Category of Objects by Transferring Grasps from Only One Labeled Instance

Hongtao Wen, Jianhang Yan, Wanli Peng, Yi Sun

ECCV 2022 · 3D vision › Human pose estimation · approach: Category-level pose estimation · 26 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 dense-correspondence grasp transfer instead of direct mapping (ablation)Direct Mapping · Simulation transfer success rate, GT pose (average)transfer success rate (%): 57.92 -> 71.95Improved
add grasp pose refinement module to grasp transfer (ablation)Grasp Transfer · Simulation transfer success rate, GT pose (average)transfer success rate (%): 71.95 -> 78.84Improved
add refinement module to direct mapping (ablation)Direct Mapping · Simulation transfer success rate, GT pose (average)transfer success rate (%): 57.92 -> 57.98No clear effect
add refinement to grasp transfer on bowl (ablation)Grasp Transfer · Simulation transfer success rate, GT pose (Bowl)transfer success rate (%): 46.24 -> 61.52Improved
add refinement to grasp transfer on mug (ablation)Grasp Transfer · Simulation transfer success rate, GT pose (Mug)transfer success rate (%): 82.67 -> 87.77Improved
add refinement to grasp transfer on bottle (ablation)Grasp Transfer · Simulation transfer success rate, GT pose (Bottle)transfer success rate (%): 86.93 -> 87.24No clear effect
add refinement to direct mapping on mug (ablation)Direct Mapping · Simulation transfer success rate, GT pose (Mug)transfer success rate (%): 60.68 -> 59.26Hurt
use TransGrasp instead of GPD in simulationGPD · Simulation grasp success, estimated pose (average)grasp success rate (%): 51.08 -> 82.50Improved
use TransGrasp instead of 6-DOF GraspNet in simulation6-DOF GraspNet · Simulation grasp success, estimated pose (average)grasp success rate (%): 76.74 -> 82.50Improved
use TransGrasp instead of 6-DOF GraspNet on bowl6-DOF GraspNet · Simulation grasp success, estimated pose (Bowl)grasp success rate (%): 75.93 -> 72.84Hurt
use TransGrasp for inference time6-DOF GraspNet · Simulation grasp success, estimated poseinference time (s): 1.82 -> 0.53Improved
use TransGrasp instead of GPD on real robotGPD · Real robot, averagegrasp success rate (%): 37/75 -> 59/75Improved

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