AI Paper Atlas

Research map / 3D vision / Human pose estimation

Recovering 3D Hand Mesh Sequence From a Single Blurry Image: A New Dataset and Temporal Unfolding

Yeonguk Oh, JoonKyu Park, Jaeha Kim, Gyeongsik Moon, Kyoung Mu Lee

CVPR 2023 · 3D vision › Human pose estimation · approach: Human mesh recovery from images · 15 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
Train on BlurHand rather than InterHand2.6MBlurHandNet · BlurHandMPJPE (mm): 24.57 -> 16.80Improved
Add BlurHand training data instead of training only on sharp InterHand2.6MBlurHandNet · InterHand2.6MMPJPE (mm): 15.33 -> 16.12Hurt
Add Unfolder and KTFormer (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 17.89 -> 16.80Improved
Add KTFormer without Unfolder, using single-timestep features (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 17.89 -> 17.41Improved
Add Unfolder without KTFormer (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 17.89 -> 17.55Improved
Deblur the image before using Unfolder and KTFormer (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 16.80 -> 17.28Hurt
Use 5 unfolded hands instead of 3 (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 16.80 -> 16.78Improved
Add both kinematic and temporal positional embeddings (ablation)KTFormer · BlurHand, middleMPJPE (mm): 17.79 -> 16.80Improved
Add kinematic positional embedding (ablation)KTFormer · BlurHand, middleMPJPE (mm): 17.79 -> 16.92Improved
Add temporal positional embedding (ablation)KTFormer · BlurHand, middleMPJPE (mm): 17.79 -> 17.41Improved
Add temporal order-invariant loss (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 16.80 -> 17.14Hurt
Add Unfolder-driven temporal ordering (ablation)BlurHandNet · BlurHand, middleMPJPE (mm): 17.14 -> 16.80Improved

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