Research map / 3D vision / Human pose estimation
RT-GENE: Real-Time Eye Gaze Estimation in Natural Environments
Tobias Fischer, Hyung Jin Chang, Yiannis Demiris
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 |
|---|---|---|---|
| train on inpainted images instead of original images with glasses (ablation) | Proposed 4 model ensemble · RT-GENE dataset | 3D gaze error (deg): 8.6 -> 7.7 | Improved |
| train on inpainted images instead of original images with glasses (ablation) | Single eye [60] · RT-GENE dataset | 3D gaze error (deg): 14.9 -> 13.4 | Improved |
| train on inpainted images instead of original images with glasses (ablation) | Spatial weights CNN [61] · RT-GENE dataset | 3D gaze error (deg): 10.0 -> 8.7 | Improved |
| inpaint glasses region instead of leaving original images (ablation) | CLNF landmark detector · dataset images | face detection rate (%): 54.6±24.7 -> 87.7±15.6 | Improved |
| inpaint glasses region instead of leaving original images (ablation) | ERT landmark detector · dataset images | landmark error (pixel): 6.6±2.3 -> 5.1±1.3 | Improved |
| inpaint glasses region instead of uniform mean-color fill (ablation) | ERT landmark detector · dataset images | face detection rate (%): 59.7±23.0 -> 84.1±17.9 | Improved |
| inpaint glasses region instead of uniform mean-color fill (ablation) | CLNF landmark detector · dataset images | landmark error (pixel): 5.6±2.3 -> 5.3±1.8 | Improved |
| use an ensemble of 4 models instead of a single model (ablation) | Proposed network · MPII Gaze | 3D gaze error (deg): 4.8 -> 4.3 | Improved |
| use a 2 model ensemble instead of a single model (ablation) | Proposed network · MPII Gaze | 3D gaze error (deg): 4.8 -> 4.6 | Improved |
| apply ensemble to spatial weights CNN (ablation) | Spatial weights CNN [61] · MPII Gaze | 3D gaze error (deg): 4.8 -> 4.8 | No clear effect |
| apply ensemble to spatial weights CNN (ablation) | Spatial weights CNN [61] · RT-GENE dataset | 3D gaze error (deg): 8.7 -> 8.7 | No clear effect |
| train on RT-GENE rather than UT Multi-view for cross-dataset test on MPII (ablation) | Proposed ensemble · MPII Gaze (cross-dataset) | 3D gaze error (deg): 8.9 -> 7.7 | Improved |
Closest papers by meaning
- MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation (TPAMI 2017 · similarity 0.90)
- Towards High Performance Low Complexity Calibration in Appearance Based Gaze Estimation (TPAMI 2022 · similarity 0.89)
- Appearance-Based Gaze Estimation With Deep Learning: A Review and Benchmark (TPAMI 2024 · similarity 0.89)
- Enhancing 3D Gaze Estimation in the Wild using Weak Supervision with Gaze Following Labels (CVPR 2025 · similarity 0.88)
- Gaze from Origin: Learning for Generalized Gaze Estimation by Embedding the Gaze Frontalization Process (AAAI 2024 · similarity 0.87)
- GazeGene: Large-scale Synthetic Gaze Dataset with 3D Eyeball Annotations (CVPR 2025 · similarity 0.87)
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.
