Research map / Image generation / Image restoration
Blind Geometric Distortion Correction on Images Through Deep Learning
Xiaoyu Li, Bo Zhang, Pedro V. Sander, Jing Liao
Check 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 |
|---|---|---|---|
| Use model-constrained flow prediction in place of unconstrained direct flow regression (ablation) | GeoNetS versus GeoNetM w/o Clas, trained on single-type data · Six distortion types | EPE: 1.82 -> 1.46 | Improved |
| Add joint distortion-type classification (ablation) | GeoNetM trained on multi-type distortion data · Six distortion types | EPE: 3.19 -> 2.75 | Improved |
| Add Hough-transform model fitting (ablation) | GeoNetM · Six distortion types | EPE: 2.75 -> 1.94 | Improved |
| Train on multi-type rather than single-type distortion data (ablation) | GeoNetM w/o Clas · Six distortion types | EPE: 1.82 -> 3.19 | Hurt |
| Initialize the resampling search using a finite-difference flow estimate | Resampling iterative search · Images with different distortion levels | iterations to convergence: roughly 10 -> 5; 15 -> 5 | Improved |
| Use the proposed correction instead of the previous perspective correction method | Perspective correction · 30 building images | angle deviation: $3.03^{\circ}$ -> $2.81^{\circ}$ | Improved |
| Use the proposed correction instead of the input image | Perspective correction · 30 building images | angle deviation: $6.44^{\circ}$ -> $2.81^{\circ}$ | Improved |
| Use the proposed method rather than specialized lens-distortion methods | Lens distortion correction · 50 randomly chosen images | number of images within relative-error thresholds | Improved |
Closest papers by meaning
- Geometry Grounding: Elevating Blind Distortion Correction with 3D Structural Priors (ECCV 2026 · similarity 0.88)
- Feature Quantization for Defending Against Distortion of Images (CVPR 2018 · similarity 0.85)
- Wide-Angle Image Rectification: A Survey (IJCV 2022 · similarity 0.85)
- Deep Generative Model based Rate-Distortion for Image Downscaling Assessment (CVPR 2024 · similarity 0.83)
- Distilling Spatially-Heterogeneous Distortion Perception for Blind Image Quality Assessment (CVPR 2025 · similarity 0.83)
- Boosting Image Quality Assessment through Efficient Transformer Adaptation with Local Feature Enhancement (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.
