PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
May 22, 2026International Journal of Computer VisionOpen Access

Leveraging Synthetic Data for Enhancing Egocentric Hand-Object Interaction Detection

View Full Paper
Ask AI
Bookmark
Share

Authors

RLRosario LeonardiAFAntonino FurnariFRFrancesco Ragusa

Discussion

Loading...

Member takes

Overview

Randomized trial explores synthetic data to improve hand-object interaction detection in egocentric images, suggesting significant gains over real data.

Key Points

  • This work aims to improve hand-object interaction detection by leveraging synthetic data in egocentric images.
  • Utilized VISOR, EgoHOS, and ENIGMA-51 datasets for comparative analysis.
  • Trained models using 10% real labeled data combined with synthetic data.
  • Developed a data generation pipeline and released the HOI-Synth benchmark for synthetic hand-object interaction data.
  • Achieved +5.67% improvement in Overall AP on VISOR with synthetic data.
  • Achieved +8.24% improvement in Overall AP on EgoHOS with synthetic data.
  • Achieved +11.69% improvement in Overall AP on ENIGMA-51 with synthetic data.

Cite This Study

Leonardi et al. (2026) studied this question.

synapsesocial.com/papers/6a0ff3c0d674f7c03778c941https://doi.org/10.1007/s11263-026-02838-8
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Exploiting multimodal synthetic data for egocentric human-object interaction detection in an industrial scenario2024 · 22 citations
  2. 2GenHOI: Generalizing Text-driven 4D Human-Object Interaction Synthesis for Unseen Objects2025
  3. 3Synthetic data enables faster annotation and robust segmentation for multi-object grasping in clutter2024 · 1 citations
  4. 4Gaze-guided Hand-Object Interaction Synthesis: Benchmark and Method2024 · 1 citations
  5. 5OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model2025