Research map / Detection and segmentation / CNN features
Boundary Flow: A Siamese Network That Predicts Boundary Motion Without Training on Motion
Peng Lei, Fuxin Li, Sinisa Todorovic
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 |
|---|---|---|---|
| Use the Siamese FCSN for boundary detection instead of CEDN | Boundary detection · VSB100 | ODS: 0.563 -> 0.597 | Improved |
| Use the Siamese FCSN for boundary detection instead of CEDN | Boundary detection · VSB100 | AP: 0.547 -> 0.566 | Improved |
| Fine-tune on BSDS500 and VSB100 training sets | FCSN · VSB100 | ODS: 0.597 -> 0.698 | Improved |
| Use edgelet-based matching instead of FLANN | Boundary flow estimation · Sintel training | EPE: $23.158$ -> $9.856$ | Improved |
| Use edgelet-based matching instead of RANSAC | Boundary flow estimation · Sintel training | EPE: $20.874$ -> $9.856$ | Improved |
| Add boundary-flow matches to CPM-Flow input | CPM-Flow · Sintel final test | EPE all: $5.960$ -> 5.645 | Improved |
| Add boundary-flow matches to CPM-Flow input | CPM-Flow · Sintel final test | EPE unmatched: $30.177$ -> \textbf{29.362} | Improved |
| Add boundary-flow matches to CPM-Flow input | CPM-Flow · Sintel final test | EPE matched: $2.990$ -> $2.737$ | Hurt |
Closest papers by meaning
- Temporal Perceiver: A General Architecture for Arbitrary Boundary Detection (TPAMI 2023 · similarity 0.88)
- Learning on the Edge: Investigating Boundary Filters in CNNs (IJCV 2019 · similarity 0.88)
- Deep Scene Flow Learning: From 2D Images to 3D Point Clouds (TPAMI 2023 · similarity 0.87)
- DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method (NeurIPS 2025 · similarity 0.86)
- FlowFeat: Pixel-Dense Embedding of Motion Profiles (NeurIPS 2025 · similarity 0.86)
- Injecting Frame-Event Complementary Fusion into Diffusion for Optical Flow in Challenging Scenes (NeurIPS 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.
