Research map / 3D vision / Face and shape
RTGaze: Real-Time 3D-Aware Gaze Redirection from a Single Image
Hengfei Wang, Zhongqun Zhang, Yihua Cheng, Hyung Jin Chang
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 |
|---|---|---|---|
| inject gaze prompt into high-frequency instead of low-frequency feature (ablation) | RTGaze · ETH-XGaze | FID: 67.298 -> 38.346 | Improved |
| inject gaze prompt into high-frequency instead of low-frequency feature (ablation) | RTGaze · ETH-XGaze | gaze error: 18.973 -> 9.047 | Improved |
| inject gaze prompt into high-frequency instead of low-frequency feature (ablation) | RTGaze · ETH-XGaze | ID: 38.517 -> 60.708 | Improved |
| inject gaze prompt into both low and high-frequency features (ablation) | RTGaze · ETH-XGaze | training convergence | Hurt |
| add perceptual loss L_P to L_R (ablation) | RTGaze · ETH-XGaze | FID: 101.053 -> 54.682 | Improved |
| add perceptual loss L_P to L_R (ablation) | RTGaze · ETH-XGaze | gaze error: 9.332 -> 10.911 | Hurt |
| add 3D face prior distillation L_D (ablation) | RTGaze · ETH-XGaze | FID: 54.682 -> 38.346 | Improved |
| add 3D face prior distillation L_D (ablation) | RTGaze · ETH-XGaze | ID: 52.518 -> 60.708 | Improved |
| add 3D face prior distillation L_D (ablation) | RTGaze · ETH-XGaze | gaze error: 10.911 -> 9.047 | Improved |
| use RTGaze instead of GazeNeRF | GazeNeRF · ETH-XGaze | gaze error: 6.944 -> 9.047 | Hurt |
| use RTGaze instead of GazeNeRF | GazeNeRF · ETH-XGaze | SSIM: 0.733 -> 0.715 | No clear effect |
| use RTGaze instead of GazeNeRF | GazeNeRF · ETH-XGaze | total inference time (s): 60.060s -> 0.061s | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- GazeNeRF — GazeNeRF: 3D-Aware Gaze Redirection With Neural Radiance Fields (CVPR 2023) · better on 9 of 10 shared results
- HeadNeRF — HeadNeRF: A Real-Time NeRF-Based Parametric Head Model (CVPR 2022) · better on 6 of 7 shared results
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
- GazeNeRF: 3D-Aware Gaze Redirection With Neural Radiance Fields (CVPR 2023 · similarity 0.91)
- GazeGaussian: High-Fidelity Gaze Redirection with 3D Gaussian Splatting (ICCV 2025 · similarity 0.91)
- What we need is explicit controllability: Training 3D gaze estimator using only facial images (ICCV 2025 · similarity 0.89)
- Self-Learning Transformations for Improving Gaze and Head Redirection (NeurIPS 2020 · similarity 0.89)
- GA3CE: Unconstrained 3D Gaze Estimation with Gaze-Aware 3D Context Encoding (CVPR 2025 · similarity 0.87)
- De^2Gaze: Deformable and Decoupled Representation Learning for 3D 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.
