Research map / Image generation / Image restoration
RecRecNet: Rectangling Rectified Wide-Angle Images by Thin-Plate Spline Model and DoF-based Curriculum Learning
Kang Liao, Lang Nie, Chunyu Lin, Zishuo Zheng, Yao Zhao
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 |
|---|---|---|---|
| apply RecRecNet rectangling to rectified wide-angle images before perception | Mask R-CNN on rectified images · Rectangling test set (500 images) | AP: 34.7 -> 41.3 (vs He et al.) | Improved |
| use RecRecNet instead of He et al. for rectangling | RecRecNet · Rectangling test set (500 images) | PSNR: 15.36 -> 18.68 | Improved |
| use RecRecNet instead of cropping the rectified image | Mask R-CNN on rectified images · Rectangling test set (500 images) | mIoU: 18.6 -> 37.8 | Improved |
| increase TPS control point grid from 3x3 to 9x9 (ablation) | RecRecNet · Rectangling test set (500 images) | PSNR: 17.96 -> 18.68 | Improved |
| increase control point grid from 9x9 to 11x11 (ablation) | RecRecNet · Rectangling test set (500 images) | PSNR: 18.68 -> 18.64 | No clear effect |
| add DoF-based curriculum learning (4-DoF to 8-DoF) before rectangling (ablation) | RecRecNet · Rectangling training and validation | convergence and overfitting: vanilla needs about 40 more epochs to converge | Improved |
| choose curriculum 4-DoF to 8-DoF (C3) over other curricula (ablation) | RecRecNet · Rectangling structure recovery | boundary-aware structure recovery | Improved |
| increase control points from 3x3 to 9x9 (ablation) | RecRecNet · Rectangling test set (500 images) | FID: 25.00 -> 19.01 | Improved |
Closest papers by meaning
- Lifting the Structural Morphing for Wide-Angle Images Rectification: Unified Content and Boundary Modeling (ICCV 2025 · similarity 0.91)
- Towards Complete Scene and Regular Shape for Distortion Rectification by Curve-Aware Extrapolation (ICCV 2021 · similarity 0.88)
- Rectification Reimagined: A Unified Mamba Model for Image Correction and Rectangling with Prompts (AAAI 2026 · similarity 0.88)
- Multi-Level Curriculum for Training a Distortion-Aware Barrel Distortion Rectification Model (ICCV 2021 · similarity 0.87)
- Rectification-specific Supervision and Constrained Estimator for Online Stereo Rectification (CVPR 2025 · similarity 0.85)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · similarity 0.85)
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.
