Research map / Image generation / Editing and style transfer
DeepWarp: Photorealistic Image Resynthesis for Gaze Manipulation
Yaroslav Ganin, Daniil Kononenko, Diana Sungatullina, Victor Lempitsky
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 lightness correction module to coarse-to-fine warping (ablation) | CFW · Gaze redirection user study (16 subjects) | correctly guessed (out of 40), mean: 28.8 -> 25.3 | Improved |
| add lightness correction module to coarse-to-fine warping (ablation) | CFW · Gaze redirection user study (16 subjects) | correctly guessed within 2 s, mean: 11.7 -> 8.0 | Improved |
| add lightness correction module to coarse-to-fine warping (ablation) | CFW · Gaze redirection user study (16 subjects) | correctly guessed within 1 s, mean: 1.6 -> 1.1 | Improved |
| add lightness correction module to coarse-to-fine warping (ablation) | CFW · Gaze redirection user study (16 subjects) | mean time to make a guess (s): 3.60 -> 3.96 | Improved |
| replace single-scale warping with coarse-to-fine warping (ablation) | Single Scale · Gaze redirection user study (16 subjects) | correctly guessed (out of 40), mean: 33.8 -> 28.8 | Improved |
| replace single-scale warping with coarse-to-fine warping (ablation) | Single Scale · Gaze redirection user study (16 subjects) | correctly guessed within 2 s, mean: 21.1 -> 11.7 | Improved |
| replace random forest with single-scale deep warping network | Random Forest · Gaze redirection user study (16 subjects) | correctly guessed (out of 40), mean: 36.1 -> 33.8 | Improved |
| replace random forest with full model CFW+LCM | Random Forest · Gaze redirection user study (16 subjects) | mean time to make a guess (s): 1.89 -> 3.96 | Improved |
| add total variation penalty on residual flow (ablation) | CFW · Gaze redirection test set | l2 loss | Hurt |
| add extra weight on eye-region pixels in the loss (ablation) | CFW · Gaze redirection test set | l2 loss | No clear effect |
| add adversarial loss (ablation) | CFW · Gaze redirection | image quality | No clear effect |
| add more palette colors for admixing in lightness module (ablation) | CFW+LCM · Gaze redirection | photorealism | No clear effect |
Closest papers by meaning
- Photorealistic Monocular Gaze Redirection Using Machine Learning (TPAMI 2017 · similarity 0.90)
- Photo-Realistic Monocular Gaze Redirection Using Generative Adversarial Networks (ICCV 2019 · similarity 0.89)
- Self-Learning Transformations for Improving Gaze and Head Redirection (NeurIPS 2020 · similarity 0.89)
- GazeNeRF: 3D-Aware Gaze Redirection With Neural Radiance Fields (CVPR 2023 · similarity 0.88)
- RTGaze: Real-Time 3D-Aware Gaze Redirection from a Single Image (AAAI 2026 · similarity 0.87)
- Neural Implicit Morphing of Face Images (CVPR 2024 · similarity 0.84)
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.
