| Benchmark | Model | Rank | Results |
|---|---|---|---|
| text-spotting-on-inverse-text | DeepSolo (ViTAEv2-S, TextOCR) | #1 | F-measure (%) - No Lexicon: 68.8F-measure (%) - Full Lexicon: 75.8 |
| text-spotting-on-inverse-text | DeepSolo (ResNet-50, TextOCR) | #2 | F-measure (%) - No Lexicon: 64.6F-measure (%) - Full Lexicon: 71.2 |
| text-spotting-on-inverse-text | DeepSolo (ResNet-50) | #4 | F-measure (%) - No Lexicon: 48.5F-measure (%) - Full Lexicon: 53.9 |
| text-spotting-on-scut-ctw1500 | DeepSolo (ResNet-50) | #3 | F-measure (%) - No Lexicon: 64.2F-Measure (%) - Full Lexicon: 81.4 |