Model card
EK-Net.
Zhu et al.open-sourceUnknown paramsResNet-18 + Expand Kernel Distance
Real-time scene text detector with expand kernel distance. 40.13 FPS on CTW1500.
§ 01 · Benchmarks
Every benchmark EK-Net has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | ICDAR 2015 | Computer Vision · Scene Text Detection | precision | 92.0% | #4 | — | source ↗ |
| 02 | scut-ctw1500 | Computer Vision · Optical Character Recognition | f-measure | 85.8% | #8 | 2024-01-22 | source ↗ |
| 03 | scut-ctw1500 | Computer Vision · Optical Character Recognition | precision | 87.8% | #8 | 2024-01-22 | source ↗ |
| 04 | scut-ctw1500 | Computer Vision · Optical Character Recognition | recall | 83.7% | #9 | 2024-01-22 | source ↗ |
| 05 | ICDAR 2015 | Computer Vision · Scene Text Detection | f-measure | 85.7% | #24 | — | source ↗ |
| 06 | ICDAR 2015 | Computer Vision · Scene Text Detection | recall | 80.2% | #29 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Papers
1 paper with results for EK-Net.
- 2024-01-22· Computer Vision· 3 results
EK-Net: Real-time Scene Text Detection with Expand Kernel Distance
§ 05 · Sources & freshness
Where these numbers come from.
arxiv
6
results
3 of 6 rows marked verified.