End-to-end scene text spotting on SCUT-CTW1500
64.2HMean (None)DeepSolo
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DeepSololen=50, Backbone=Res-502023.05 | 64.2 | 81.4 | 20 | |
| SPTSlen=1002023.05 | 63.6 | 83.8 | 0.8 | |
| DeepSololen=50, Backbone=Res-50, Pre-training=Synth150K2023.05 | 63.2 | 80 | 20 | |
| ABINet++Character-level annotations=false, Multi-scale=true2022.11 | 61.5 | 81.2 | — | |
| ABINet++Character-level annotations=false, Multi-scale=false2022.11 | 60.2 | 80.3 | 10.6 | |
| ABINet++len=1002023.05 | 60.2 | 80.3 | — | |
| DeepSololen=25, Backbone=Res-50, Pre-training=Synth150K2023.05 | 60.1 | 78.4 | 20 | |
| TPSNetlen=100, Backbone=Res-50-DCN2023.05 | 59.7 | 79.2 | — | |
| MANGOYear=2021, Character-level annotations=false, Multi-scale=false2022.11 | 58.9 | 78.7 | 4.3 | |
| MANGOlen=252023.05 | 58.9 | 78.7 | 8.4 | |
| ABCNet v2Year=2021, Character-level annotations=false, Multi-scale=false2022.11 | 57.5 | 77.2 | 10 | |
| ABCNet v2len=1002023.05 | 57.5 | 77.2 | 10 | |
| TextPerceptron2023.05 | 57 | — | — | |
| TESTRlen=1002023.05 | 56 | 81.5 | 15.9 | |
| SwinTSBackbone=Swin-T2023.05 | 51.8 | 77 | — | |
| ABCNetYear=2020, Character-level annotations=false, Multi-scale=false2022.11 | 45.2 | 74.1 | 17.9 | |
| Feng et al.Year=2021, Character-level annotations=false, Multi-scale=false2022.11 | 42.2 | 74.9 | 7.2 | |
| TextDragonYear=2019, Character-level annotations=false, Multi-scale=false2022.11 | 39.7 | 72.4 | 2.6 | |
| TextDragon2023.05 | 39.7 | 72.4 | — | |
| FOTSYear=2018, Character-level annotations=false, Multi-scale=false2022.11 | 21.1 | 39.7 | — |