Text Detection on MLT 2019 (test)
75.5F-scoreCRAFTS
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CRAFTSSource=Original paper [3]2021.03 | 75.5 | 81.7 | 70.1 | — | — | — | — | — | — | — | — | |
| Multiplexed TextSpotter2021.03 | 72.66 | 85.53 | 63.16 | 60.46 | 51.75 | 73.55 | 43.86 | 42.43 | 57.15 | 40.27 | 51.95 | |
| DBNet++Backbone=ResNet-50, FPS=102022.02 | 71.4 | 78.6 | 65.4 | — | — | — | — | — | — | — | — | |
| Single-head TextSpotter2021.03 | 71.1 | 83.75 | 61.76 | 58.76 | 51.12 | 73.56 | 40.41 | 41.22 | 56.54 | 39.68 | 49 | |
| CRAFTSSource=Official ICDAR19 MLT website2021.03 | 70.86 | 81.42 | 62.73 | 56.63 | 43.97 | 72.49 | 37.2 | 42.1 | 54.05 | 38.5 | 53.5 | |
| DBNetBackbone=ResNet-50, FPS=102022.02 | 70.4 | 78.3 | 64 | — | — | — | — | — | — | — | — | |
| RRPN2021.03 | 69.56 | 77.71 | 62.95 | 58.07 | 35.88 | 68.01 | 33.31 | 36.11 | 45.06 | 28.78 | 40 | |
| DBNet++Backbone=ResNet-18, FPS=182022.02 | 68.2 | 77.5 | 61 | — | — | — | — | — | — | — | — | |
| CRAFTS*Backbone=Not specified, Notes=Used character-level annotations and integrated a recognition model2022.02 | 68.1 | 79.5 | 59.6 | — | — | — | — | — | — | — | — | |
| DBNetBackbone=ResNet-18, FPS=192022.02 | 66.9 | 75.3 | 60.2 | — | — | — | — | — | — | — | — | |
| PSENet2021.03 | 65.83 | 73.52 | 59.59 | 52.73 | 43.96 | 65.77 | 38.47 | 34.47 | 51.73 | 34.04 | 47.19 | |
| PSENetBackbone=Not specified2022.02 | 65.8 | 73.5 | 59.6 | — | — | — | — | — | — | — | — |