Research map / 3D vision / 3D generation
WarpNet: Weakly Supervised Matching for Single-View Reconstruction
Angjoo Kanazawa, David W. Jacobs, Manmohan Chandraker
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 |
|---|---|---|---|
| Add WarpNet's learned spatial prior to appearance-based matching | VGG-M conv4 matching · CUB-200-2011 test pairs | AP: +13.6% | Improved |
| Train with exemplar-TPS transformations instead of affine spatial transformations (ablation) | WarpNet · CUB-200-2011 test pairs | matching precision and recall | Improved |
| Use affine spatial transformations to create the artificial training data (ablation) | WarpNet · CUB-200-2011 test pairs | PCK at 85% precision | Hurt |
| Use co-segmentation outputs instead of ground-truth segmentation (ablation) | WarpNet · CUB-200-2011 test pairs | AP improvement over VGG-M: +10.8% | Improved |
| Replace SIFT features with VGG features in DSP | DSP · CUB-200-2011 test pairs | final recall: around 5% | Improved |
| Use WarpNet instead of DSP as the spatial prior | Matching with a spatial prior · CUB-200-2011 test pairs | matching performance | Improved |
Closest papers by meaning
- Warp Consistency for Unsupervised Learning of Dense Correspondences (ICCV 2021 · similarity 0.88)
- Probabilistic Warp Consistency for Weakly-Supervised Semantic Correspondences (CVPR 2022 · similarity 0.88)
- AnchorNet: A Weakly Supervised Network to Learn Geometry-Sensitive Features for Semantic Matching (CVPR 2017 · similarity 0.87)
- ASIC: Aligning Sparse in-the-wild Image Collections (ICCV 2023 · similarity 0.87)
- Weakly Supervised Learning of Semantic Correspondence through Cascaded Online Correspondence Refinement (ICCV 2023 · similarity 0.86)
- End-to-End Weakly-Supervised Semantic Alignment (CVPR 2018 · 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.
