Model card
DBNet (ResNet-18+DCN).
Liao et al. (USTC)open-sourceUnknown paramsResNet-18 + Deformable Convolution + Differentiable Binarization
Original DBNet (AAAI 2020) with ResNet-18+DCN backbone. arXiv:1911.08947. Result from TextBPN++ paper Table IX.
§ 01 · Benchmarks
Every benchmark DBNet (ResNet-18+DCN) has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | icdar-2017-mlt | Computer Vision · Scene Text Detection | precision | 81.9% | #6 | 2022-05-11 | source ↗ |
| 02 | icdar-2017-mlt | Computer Vision · Scene Text Detection | f-measure | 71.7% | #14 | 2022-05-11 | source ↗ |
| 03 | icdar-2017-mlt | Computer Vision · Scene Text Detection | recall | 63.8% | #15 | 2022-05-11 | 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.
§ 02 · Strengths by area
Where DBNet (ResNet-18+DCN) actually performs.
§ 03 · Papers
1 paper with results for DBNet (ResNet-18+DCN).
- 2022-05-11· Computer Vision· 3 results
Arbitrary Shape Text Detection via Boundary Transformer
§ 04 · Related models
Other Liao et al. (USTC) models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
arxiv
3
results
3 of 3 rows marked verified.