Research map / 3D vision / Human pose estimation
Dynamic 3D Gaze From Afar: Deep Gaze Estimation From Temporal Eye-Head-Body Coordination
Soma Nonaka, Shohei Nobuhara, Ko Nishino
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use temporal head-body coordination to estimate gaze (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 24.1 -> 21.7 | Improved |
| Use estimated head and body uncertainty in the gaze network (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 22.1 -> 21.7 | Improved |
| Include body direction as an input (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 24.1 -> 21.7 | Improved |
| Include head direction as an input (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 24.9 -> 21.7 | Improved |
| Add opportunistic eye-appearance integration (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 21.7 -> 20.4 | Improved |
| Train the head and body network on AGORA rather than GAFA (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 21.7 -> 27.3 | Hurt |
| Combine monocular predictions using uncertainty-weighted multiview averaging (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 21.6 -> 20.8 | Improved |
| Use multiview observations and 3D body velocity (ablation) | Ours · GAFA test set, all scenes | 3D MAE: 21.6 -> 18.9 | Improved |
| Train the head and body network on AGORA and the gaze network on GAFA | Ours · MoDiPro | 2D MAE: 46.3 -> 25.6 | Improved |
| Use temporal head-body coordination instead of static facial keypoints | Ours versus Dias et al. · MoDiPro | 2D MAE: 32.1 -> 25.6 | Improved |
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.
- Gaze360 — Gaze360: Physically Unconstrained Gaze Estimation in the Wild (ICCV 2019) · better on 1 of 1 shared result
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
- Gaze360: Physically Unconstrained Gaze Estimation in the Wild (ICCV 2019 · similarity 0.91)
- 3DGazeNet: Generalizing Gaze Estimation with Weak Supervision from Synthetic Views (ECCV 2024 · similarity 0.90)
- GA3CE: Unconstrained 3D Gaze Estimation with Gaze-Aware 3D Context Encoding (CVPR 2025 · similarity 0.90)
- Appearance-Based Gaze Estimation With Deep Learning: A Review and Benchmark (TPAMI 2024 · similarity 0.90)
- Multiple-gaze geometry: Inferring novel 3D locations from gazes observed in monocular video (ECCV 2018 · similarity 0.90)
- Enhancing 3D Gaze Estimation in the Wild using Weak Supervision with Gaze Following Labels (CVPR 2025 · similarity 0.90)
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.
