Research map / Probabilistic and scientific modeling / Molecules and proteins
DCTM: Discrete-Continuous Transformation Matching for Semantic Flow
Seungryong Kim, Dongbo Min, Stephen Lin, Kwanghoon Sohn
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 |
|---|---|---|---|
| add continuous regularization to the discrete optimization (ablation) | DCTM · Taniai benchmark (T=5px) average | flow accuracy: 0.682 -> 0.740 | Improved |
| add coarse-to-fine scheme (ablation) | DCTM · Taniai benchmark (T=5px) average | flow accuracy: 0.698 -> 0.740 | Improved |
| use affine-FCSS instead of DAISY descriptor (ablation) | DCTM · Taniai benchmark (T=5px) average | flow accuracy: 0.566 -> 0.740 | Improved |
| use affine-FCSS instead of VGG descriptor (ablation) | DCTM · Taniai benchmark (T=5px) average | flow accuracy: 0.630 -> 0.740 | Improved |
| add continuous regularization (ablation) | DCTM · Taniai benchmark (T=5px) PASCAL group | flow accuracy: 0.559 -> 0.610 | Improved |
| add coarse-to-fine scheme (ablation) | DCTM · Taniai benchmark (T=5px) JODS group | flow accuracy: 0.684 -> 0.721 | Improved |
| add geometric (affine) modeling with DCTM over SF w/FCSS | SF w/FCSS · Taniai benchmark (T=5px) average | flow accuracy: 0.660 -> 0.740 | Improved |
| replace SF w/FCSS with DCTM | SF w/FCSS · Proposal Flow (alpha=0.05) | PCK: 0.354 -> 0.381 | Improved |
| replace Proposal Flow with DCTM | Proposal Flow · Proposal Flow (alpha=0.1) | PCK: 0.568 -> 0.610 | Improved |
| replace SF w/FCSS with DCTM | SF w/FCSS · PASCAL-VOC Parts | weighted IoU: 0.44 -> 0.48 | Improved |
Closest papers by meaning
- Discrete-Continuous Transformation Matching for Dense Semantic Correspondence (TPAMI 2018 · similarity 0.98)
- Guided Semantic Flow (ECCV 2020 · similarity 0.90)
- Pixel-level Semantic Correspondence through Layout-aware Representation Learning and Multi-scale Matching Integration (CVPR 2024 · similarity 0.89)
- Deep Semantic Feature Matching (CVPR 2017 · similarity 0.88)
- Learning Semantic Correspondence Exploiting an Object-Level Prior (TPAMI 2020 · similarity 0.88)
- SFNet: Learning Object-Aware Semantic Correspondence (CVPR 2019 · similarity 0.88)
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.
