AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Human Hands As Probes for Interactive Object Understanding

Mohit Goyal, Sahil Modi, Rishabh Goyal, Saurabh Gupta

CVPR 2022 · Detection and segmentation › Open-set and few-shot detection · approach: Hand-object and articulated interaction · 29 citations (OpenAlex)

Check 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 hand appearance and motion consistency to temporal contrastive learningTSC object features · EPIC-STATES test, all objects, 100% training datamAP: 84.2 $\pm$ 1.0 -> 84.8 $\pm$ 0.4Improved
Add hand appearance and motion consistency to temporal contrastive learningTSC object features · EPIC-STATES test, novel objects, 12.5% training datamAP: 74.5 $\pm$ 0.9 -> 79.7 $\pm$ 0.6Improved
Use temporal contrastive learning on object tracks instead of ImageNet featuresObject state classification · EPIC-STATES test, all objects, 100% training datamAP: 83.1 $\pm$ 0.0 -> 84.2 $\pm$ 1.0Improved
Predict interaction sites from masked surrounding context without segmentation-mask trainingACP · EPIC-ROI test, overall, 0% slackAP: 55.7 -> 57.4Improved
Combine ACP predictions with Mask R-CNNEPIC-ROI prediction · EPIC-ROI test, overall, 0% slackAP: 66.6 -> 68.7Improved
Predict grasp affordances using hand-derived supervisionACP · YCB novel-object test, 7 grasp classesmAP: 30.2% -> 38.1%Improved
Add temporal self-supervision on EPIC hand tracks to adapt the grasp classifier (ablation)ACP grasp prediction · YCB novel-object test, 7 grasp classesmAP: 34.3% -> 38.1%Improved
Add hand appearance to TSC (ablation)TSC object features · EPIC-STATES validation, novel objects, 12.5% training datamAP: 72.3 $\pm$ 1.3 -> 73.8 $\pm$ 0.8Improved
Add hand motion to TSC (ablation)TSC object features · EPIC-STATES validation, novel objects, 12.5% training datamAP: 72.3 $\pm$ 1.3 -> 74.6 $\pm$ 1.6Improved
Add both hand motion and appearance to TSC (ablation)TSC object features · EPIC-STATES validation, novel objects, 12.5% training datamAP: 72.3 $\pm$ 1.3 -> 75.1 $\pm$ 0.4Improved
Use detected object-of-interaction tracks with hand-context tracking rather than no tracking (ablation)TSC · EPIC-STATES validationmAP: 79.2 -> 81.2Improved
Replace predicted object tracks with ground-truth object-of-interaction tracks (ablation)TSC · EPIC-STATES validationmAP: 81.2 -> 83.5Improved

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