Research map / 3D vision / Face and shape
Neural Contours: Learning to Draw Lines From 3D Shapes
Difan Liu, Mohamed Nabail, Aaron Hertzmann, Evangelos Kalogerakis
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 |
|---|---|---|---|
| Combine the differentiable geometry branch with the image translation branch (ablation) | Neural Contours · Cole et al. drawings | F1 | Improved |
| Use multiple smoothed shaded images rather than a single shaded and depth image (ablation) | Image translation branch · Cole et al. drawings and new test dataset | F1 | Improved |
| Add rendered curvature maps to the image translation inputs (ablation) | Image translation branch · Line drawing benchmarks | F1 | No clear effect |
| Apply a soft threshold to the image translation output (ablation) | Neural Contours · Line drawing generation | drawing quality | No clear effect |
| Predict geometric line parameters with a feed-forward network instead of test-time optimization (ablation) | Neural Contours · Cole et al. and new test datasets | F1 | Hurt |
| Add GAN-based losses to image translation training (ablation) | Image translation branch · Line drawing generation | line selection | No clear effect |
| Use Neural Contours instead of AR-rtsc | Line drawing generation · Cole et al. drawings | F1: 54.1 -> 60.6 | Improved |
| Use Neural Contours instead of AR-rtsc | Line drawing generation · Cole et al. most consistent drawings | F1: 63.3 -> 67.6 | Improved |
| Use Neural Contours instead of AR-rtsc | Line drawing generation · New test dataset | F1: 69.9 -> 74.6 | Improved |
| Use Neural Contours instead of the best alternative in the perceptual study | Line drawing generation · MTurk perceptual evaluation | votes | Improved |
Closest papers by meaning
- Neural Strokes: Stylized Line Drawing of 3D Shapes (ICCV 2021 · similarity 0.93)
- Learning To Generate Line Drawings That Convey Geometry and Semantics (CVPR 2022 · similarity 0.88)
- Learning Manifold Patch-Based Representations of Man-Made Shapes (ICLR 2021 · similarity 0.87)
- A Neural Representation of Sketch Drawings (ICLR 2018 · similarity 0.86)
- Neural 3D Strokes: Creating Stylized 3D Scenes with Vectorized 3D Strokes (CVPR 2024 · similarity 0.86)
- SketchINR: A First Look into Sketches as Implicit Neural Representations (CVPR 2024 · 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.
