Research map / Detection and segmentation / Open-set and few-shot detection
Tracking Gaze and Visual Focus of Attention of People Involved in Social Interaction
Benoît Massé, Silèye Ba, Radu Horaud
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use proposed joint gaze-VFOA switching model instead of Ba & Odobez | Gaze/VFOA tracker · Vernissage (Vicon data) | mean FRR (%): 54.5 -> 62.6 | Improved |
| use proposed model instead of Ba & Odobez | Gaze/VFOA tracker · Vernissage (RGB data) | mean FRR (%): 39.0 -> 54.7 | Improved |
| use proposed model instead of Ba & Odobez with same head-pose input, excluding rec 26 | Gaze/VFOA tracker · Vernissage (Vicon data) | mean FRR (%): 56.5 -> 64.7 | Improved |
| use proposed model vs Sheikhi, which uses speaker and object context | Gaze/VFOA tracker · Vernissage (Vicon data) | mean FRR (%): 66.6 -> 64.7 | Hurt |
| use proposed model vs Sheikhi on RGB data | Gaze/VFOA tracker · Vernissage (RGB data) | mean FRR (%): 62.4 -> 54.7 | Hurt |
| use proposed model on hard recording 18 left person | Gaze/VFOA tracker · Vernissage (Vicon data, rec 18 left-p) | FRR (%): 37.5 -> 62.0 | Improved |
| use proposed model on recording 27 right person with RGB data | Gaze/VFOA tracker · Vernissage (RGB data, rec 27 right-p) | FRR (%): 17.1 -> 55.1 | Improved |
| use proposed model on recording 30 left person with RGB data | Gaze/VFOA tracker · Vernissage (RGB data, rec 30 left-p) | FRR (%): 61.6 -> 54.7 | Hurt |
| use proposed model on recording 10 right person with Vicon data | Gaze/VFOA tracker · Vernissage (Vicon data, rec 10 right-p) | FRR (%): 74.4 -> 65.0 | Hurt |
| replace Vicon head pose with RGB-estimated head pose | Gaze/VFOA tracker · Vernissage | mean FRR (%): 62.6 -> 54.7 | Hurt |
| use proposed model with coarse head orientation vs Ba & Odobez | Gaze/VFOA tracker · LAEO (coarse head orientation) | shot recognition rate: 0.535 -> 0.727 | Improved |
| use proposed model with fine head orientation vs Ba & Odobez | Gaze/VFOA tracker · LAEO (fine head orientation) | shot recognition rate: 0.363 -> 0.479 | Improved |
Closest papers by meaning
- Eyes on VLM: Benchmarking Gaze Following and Social Gaze Prediction in Vision Language Models (NeurIPS 2026 · similarity 0.89)
- MTGS: A Novel Framework for Multi-Person Temporal Gaze Following and Social Gaze Prediction (NeurIPS 2024 · similarity 0.88)
- Learning from Observer Gaze: Zero-Shot Attention Prediction Oriented by Human-Object Interaction Recognition (CVPR 2024 · similarity 0.87)
- EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics (AAAI 2025 · similarity 0.86)
- Gaze-VLM: Bridging Gaze and VLMs through Attention Regularization for Egocentric Understanding (NeurIPS 2025 · similarity 0.85)
- Toward Semantic Gaze Target Detection (NeurIPS 2024 · similarity 0.85)
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.
