Word Spotting on ICDAR 2015
87.8Strong ScoreDeepSolo
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DeepSoloBackbone=ViTAEv2-S, External Data=Synth150K+MLT17+IC13+Total-Text+TextOCR2023.05 | 87.8 | 84.5 | 80 | — | |
| DeepSoloBackbone=Res-50, External Data=Synth150K+MLT17+IC13+Total-Text+TextOCR2023.05 | 87.3 | 83.8 | 79.5 | — | |
| FOTS MSmulti-scale=true2018.01 | 87.01 | 82.39 | 67.97 | — | |
| GLASSExternal Data=Synth800K2023.05 | 86.8 | 82.5 | 78.8 | — | |
| DeepSoloBackbone=Res-50, External Data=Synth150K+MLT17+IC13+Total-Text2023.05 | 86.3 | 82.3 | 77.3 | — | |
| TextDragonFPS=2.62020.07 | 86.2 | 81.6 | 68 | — | |
| CharNetExternal Data=Synth800K2023.05 | 86.2 | 81.6 | 68 | — | |
| ABCNet v2External Data=Synth150K+MLT172023.05 | 85.2 | 81.1 | 74.6 | — | |
| He et al.2020.07 | 85 | 80 | 65 | — | |
| He et al.Irregular text support=false2020.02 | 85 | 80 | 65 | — | |
| TTSExternal Data=Synth800K+COCO-Text+IC13+Total-Text+SCUT2023.05 | 85 | 81.5 | 77.3 | — | |
| FOTSIrregular text support=false2020.02 | 84.7 | 79.3 | 63.3 | — | |
| FOTS2018.01 | 84.68 | 79.32 | 63.29 | — | |
| Text PerceptronMode=End-to-end, Irregular text support=false2020.02 | 84.1 | 79.4 | 67.9 | — | |
| Mask TextSpotter v3Input Resolution (short side)=1440, FPS=2.52020.07 | 83.1 | 79.1 | 75.1 | — | |
| PAN++External Data=Synth800K+COCO-Text+MLT17+Total-Text2023.05 | 83.1 | 79.1 | 75.1 | — | |
| Mask TextSpotter v2Input Resolution (short side)=1600, FPS=22020.07 | 82.4 | 78.1 | 73.6 | — | |
| TextNetIrregular text support=true2020.02 | 82.4 | 78.4 | 62.4 | — | |
| Text PerceptronMode=2-stage, Irregular text support=false2020.02 | 80.6 | 76.6 | 65.5 | — | |
| Our Two-Stagearchitecture=Two-Stage2018.01 | 80.38 | 77.66 | 58.19 | — | |
| Mask TextSpotterInput Scale=16002018.07 | 79.3 | 74.5 | 64.2 | 2.6 | |
| Mask TextSpotter v1Input Resolution (short side)=1600, FPS=2.62020.07 | 79.3 | 74.5 | 64.2 | — | |
| Mask TextSpotterIrregular text support=true2020.02 | 79.3 | 74.5 | 64.2 | — | |
| Mask TextSpotterInput Scale=10002018.07 | 77.7 | 71.3 | 58.6 | 4.8 | |
| FOTS RTreal-time=true2018.01 | 76.74 | 69.23 | 53.5 | — | |
| TextBoxes++2020.07 | 76.5 | 69 | 54.4 | — | |
| Mask TextSpotterInput Scale=7202018.07 | 71.6 | 63.9 | 51.6 | 6.9 | |
| HUST_MCLAB2018.07 | 70.6 | — | — | — | |
| HUST_MCLAB2018.01 | 70.57 | — | — | — | |
| Deep text spotter2018.07 | 58 | 53 | 51 | 9 | |
| TextSpotterIrregular text support=false2020.02 | 58 | 53 | 51 | — | |
| TextProposals+DictNet2018.01 | 56 | 52.26 | 49.73 | — | |
| TextProposals + DictNet2018.07 | 56 | 52.3 | 49.7 | 0.2 | |
| Stradvision2018.07 | 45.9 | — | — | — | |
| Stradvision-22018.01 | 45.87 | — | — | — | |
| TextSpotter2018.07 | 37 | 21 | 16 | 1 | |
| StradVision_v12018.01 | 34.65 | — | — | — | |
| NJU Text (Version3)2018.01 | 34.1 | — | — | — | |
| Beam search CUNI+S2018.01 | 23.37 | 21.07 | 18.38 | — | |
| Deep2Text-MO2018.01 | 17.58 | 17.58 | 17.58 | — | |
| Baseline OpenCV3.0 + Tesseract2018.07 | 14.7 | 12.6 | 8.4 | — | |
| Baseline OpenCV3.0+Tesseract2018.01 | 14.65 | 12.63 | 8.43 | — |