Research map / 3D vision / Depth and stereo
Practical Wide-Angle Portraits Correction With Deep Structured Models
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 basic LineNet to predict the perspective projection flow (ablation) | wide-angle portrait correction · wide-angle portrait test set | Proj LineAcc: 66.064 -> 66.745 | Helped |
| Add LAM to basic LineNet (ablation) | LineNet · wide-angle portrait test set | Proj LineAcc: 66.745 -> 66.856 | Helped |
| Add ShapeNet to LineNet (ablation) | LineNet · wide-angle portrait test set | ShapeAcc: 97.380 -> 97.458 | Helped |
| Add LAM to LineNet + ShapeNet (ablation) | LineNet + ShapeNet · wide-angle portrait test set | Corr LineAcc: 66.439 -> 66.472 | Helped |
| Add FAM to LineNet + LAM + ShapeNet (ablation) | LineNet + LAM + ShapeNet · wide-angle portrait test set | ShapeAcc: 97.472 -> 97.479 | Helped |
| Add TM to the configuration with LAM, ShapeNet, and FAM (ablation) | portrait correction network · wide-angle portrait test set | Corr LineAcc: 66.484 -> 66.575 | Helped |
| Add TM to the configuration with LAM, ShapeNet, and FAM (ablation) | portrait correction network · wide-angle portrait test set | ShapeAcc: 97.479 -> 97.485 | Helped |
| Add landmark loss to the model with LAM and ShapeNet (ablation) | portrait correction network · wide-angle portrait test set | ShapeAcc: 97.472 -> 97.473 | No clear effect |
| Add landmark loss to the model with LAM and ShapeNet (ablation) | portrait correction network · wide-angle portrait test set | Corr LineAcc: 66.472 -> 66.541 | Helped |
| Include data from multiple camera modules in training (ablation) | full model · wide-angle portrait test set | LineAcc: 66.324 -> 66.784 | Helped |
| Include data from multiple camera modules in training (ablation) | full model · wide-angle portrait test set | ShapeAcc: 97.464 -> 97.490 | Helped |
| Use the full training set instead of training without Note data (ablation) | full model · Note testset | ShapeAcc: 97.061 -> 97.115 | No clear effect |
Closest papers by meaning
- Beyond Wide-Angle Images: Structure-to-Detail Video Portrait Correction via Unsupervised Spatiotemporal Adaptation (AAAI 2026 · similarity 0.88)
- PropagationNet: Propagate Points to Curve to Learn Structure Information (CVPR 2020 · similarity 0.84)
- MaDCoW: Marginal Distortion Correction for Wide-Angle Photography with Arbitrary Objects (CVPR 2025 · similarity 0.84)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 2024 · similarity 0.83)
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.82)
- Neural Implicit Morphing of Face Images (CVPR 2024 · similarity 0.82)
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.