Research map / 3D vision / Human pose estimation
DVGaze: Dual-View Gaze Estimation
Yihua Cheng, Feng Lu
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 |
|---|---|---|---|
| add dual-view transformer to basic ResNet18 dual-view model (ablation) | DV-Gaze basic model · EVE (average) | angular error (deg): 2.97 -> 2.83 | Improved |
| add DIC blocks on top of dual-view transformer (ablation) | DV-Gaze · EVE (average) | angular error (deg): 2.83 -> 2.49 | Improved |
| add DIC blocks on top of dual-view transformer (ablation) | DV-Gaze · ETH-XGaze (average) | angular error (deg): 2.51 -> 2.46 | Improved |
| add camera pose encoding to the transformer (ablation) | DV-Gaze · ETH-XGaze (average) | angular error (deg): 2.70 -> 2.46 | Improved |
| use fused features to compensate (add to) original features instead of replacing (ablation) | DIC block · EVE (average) | angular error (deg): 2.71 -> 2.49 | Improved |
| add dual-view gaze consistency loss (ablation) | DV-Gaze · ETH-XGaze (average) | angular error (deg): 2.56 -> 2.46 | Improved |
| expand single-view training set to include dual-view images | single-view baselines · ETH-XGaze / EVE | angular error (deg): GazeTR EVE left 3.18 -> 2.92; EVE right 3.72 -> 3.81 | Mixed |
| dual-view fusion instead of single-view GazeTR | gaze estimation · ETH-XGaze / EVE | angular error (deg): 10%~30% improvement | Improved |
| dual-view fusion vs oracle best-view selection | GazeTR · ETH-XGaze | angular error (deg): 2.38 -> 2.00 | Improved |
| larger camera distance between the two views (ablation) | DV-Gaze · ETH-XGaze camera pairs | angular error improvement vs GazeTR | Improved |
Closest papers by meaning
- UVAGaze: Unsupervised 1-to-2 Views Adaptation for Gaze Estimation (AAAI 2024 · similarity 0.91)
- Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders (CVPR 2025 · similarity 0.87)
- What Do You See in Vehicle? Comprehensive Vision Solution for In-Vehicle Gaze Estimation (CVPR 2024 · similarity 0.87)
- Unsupervised Gaze Representation Learning from Multi-view Face Images (CVPR 2024 · similarity 0.86)
- De^2Gaze: Deformable and Decoupled Representation Learning for 3D Gaze Estimation (CVPR 2025 · similarity 0.86)
- FIFA: Fine-grained Inter-frame Attention for Driver's Video Gaze Estimation (CVPR 2025 · similarity 0.86)
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.
