Research map / Detection and segmentation / CNN features
MatchED: Crisp Edge Detection Using End-to-End, Matching-based Supervision
Bedrettin Cetinkaya, Sinan Kalkan, Emre Akbas
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 |
|---|---|---|---|
| integrate MatchED matching supervision | PiDiNet · BSDS (CEval) | ODS: .578 -> .800 | Improved |
| integrate MatchED | RankED · BSDS (CEval) | ODS: .631 -> .819 | Improved |
| integrate MatchED | SAUGE · BSDS (CEval) | AC: .215 -> .852 | Improved |
| integrate MatchED | DiffusionEdge · NYUD-v2 (CEval) | ODS: .727 -> .759 | Improved |
| MatchED without NMS vs standard post-processing | RankED · BSDS (SEval comparison) | ODS: -0.011 ODS | Hurt |
| MatchED without NMS vs standard post-processing | PiDiNet · BSDS (SEval comparison) | ODS: +0.011 ODS | Improved |
| MatchED without post-processing vs PiDiNet with post-processing | PiDiNet · BIPED-v2 | ODS: .892 -> .900 | Improved |
| apply NMS and thinning on top of MatchED output (ablation) | DiffusionEdge + MatchED · NYUD-v2 (SEval) | ODS | No clear effect |
| raise confidence threshold tau_c from 0.01 to 0.30 (ablation) | MatchED · BSDS (PiDiNet) | ODS: .800 -> .761 | Hurt |
| lower confidence weight alpha to 5 (ablation) | MatchED · BSDS (PiDiNet) | ODS: .800 -> .630 | Hurt |
| vary distance threshold tau_d between 2 and 8 (ablation) | MatchED · BSDS (PiDiNet) | ODS: .797 to .800 | No clear effect |
| replace NMS and 100 thinning passes with MatchED | PiDiNet · NYUD CPU runtime | time (s): 1875.57 -> 32.98 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SAUGE — SAUGE: Taming SAM for Uncertainty-Aligned Multi-Granularity Edge Detection (AAAI 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection (AAAI 2024 · similarity 0.90)
- MEMO: Human-like Crisp Edge Detection Using Masked Edge Prediction (CVPR 2026 · similarity 0.90)
- SAUGE: Taming SAM for Uncertainty-Aligned Multi-Granularity Edge Detection (AAAI 2025 · similarity 0.87)
- RankED: Addressing Imbalance and Uncertainty in Edge Detection Using Ranking-based Losses (CVPR 2024 · similarity 0.84)
- MuGE: Multiple Granularity Edge Detection (CVPR 2024 · similarity 0.84)
- STAR-Edge: Structure-aware Local Spherical Curve Representation for Thin-walled Edge Extraction from Unstructured Point Clouds (CVPR 2025 · similarity 0.83)
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.
