Scene Text Spotting on Total-Text (test)
81.7F-measure (None)Semi-DETR*
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Semi-DETR*Paradigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 81.7 | 80.3 | — | — | — | — | — | — | |
| SemiETSParadigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 81.3 | 87.3 | — | — | — | — | — | — | |
| ESTextSpotterParadigm=Supervised, Data Setting=Fully Labeled2025.04 | 80.8 | 87.1 | — | — | — | — | — | — | |
| Soft-Teacher*Paradigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 80.6 | 86.7 | — | — | — | — | — | — | |
| Mean-Teacher*Paradigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 80.1 | 87.1 | — | — | — | — | — | — | |
| GLASSParadigm=Supervised, Data Setting=Fully Labeled2025.04 | 79.9 | 86.2 | — | — | — | — | — | — | |
| DeepSoloParadigm=Supervised, Data Setting=Fully Labeled2025.04 | 79.7 | 87 | — | — | — | — | — | — | |
| Supervised baselineParadigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 79.7 | 87 | — | — | — | — | — | — | |
| Unbiased Teacher v2*Paradigm=Semi-supervised, Data Setting=Fully Labeled2025.04 | 79.6 | 87.5 | — | — | — | — | — | — | |
| CRAFTSMulti-scale inference=false2022.03 | 78.7 | — | — | — | — | — | — | — | |
| CRAFTS2023.04 | 78.7 | — | — | — | — | — | — | — | |
| UNITSModel Type=Fine-tuned2023.04 | 78.7 | 86 | — | — | — | — | — | — | |
| TTSpolyIOU threshold=0.52022.02 | 78.2 | 86.3 | — | — | — | — | — | — | |
| TTS (poly)Paradigm=Supervised, Data Setting=Fully Labeled2025.04 | 78.2 | 86.3 | — | — | — | — | — | — | |
| UNITS SharedModel Type=Unified2023.04 | 77.3 | 85 | — | — | — | — | — | — | |
| GLASS2023.04 | 76.6 | 83 | — | — | — | — | — | — | |
| TTS2023.04 | 75.6 | 84.4 | — | — | — | — | — | — | |
| MTS-V32022.02 | 75.1 | 81.8 | — | — | — | — | — | — | |
| DEERMulti-scale inference=false2022.03 | 74.84 | 81.34 | — | — | — | — | — | — | |
| DEER2023.04 | 74.8 | 83.3 | — | — | — | — | — | — | |
| SwinTextSpotter2023.04 | 74.3 | 84.1 | — | — | — | — | — | — | |
| SwinTextSpotterParadigm=Supervised, Data Setting=Fully Labeled2025.04 | 74.3 | 84.1 | — | — | — | — | — | — | |
| SPTSRepresentation=Point-based2021.12 | 74.2 | 82.4 | — | — | — | — | — | — | |
| ABCNet v2Multi-scale=true2021.05 | 73.5 | 80.7 | — | — | — | — | — | — | |
| ABCNet v2Multi-scale inference=true2022.03 | 73.5 | 80.7 | — | — | — | — | — | — | |
| TESTR-PolygonBackbone=ResNet-50, FPS=5.32022.04 | 73.3 | 83.9 | — | — | — | — | — | — | |
| TESTR2023.04 | 73.3 | 83.9 | — | — | — | — | — | — | |
| TESTRParadigm=Supervised, Data Setting=Fully Labeled2025.04 | 73.3 | 83.9 | — | — | — | — | — | — | |
| MANGOInput Size (longer side)=16002020.12 | 72.9 | 83.6 | 4.3 | — | — | — | — | — | |
| MANGO*IOU threshold=0.12022.02 | 72.9 | 83.6 | — | — | — | — | — | — | |
| MANGOMulti-scale inference=false2022.03 | 72.9 | 83.6 | — | — | — | — | — | — | |
| MANGOBackbone=ResNet-50-FPN, FPS=4.32022.04 | 72.9 | 83.6 | — | — | — | — | — | — | |
| MANGORepresentation=Bounding Box-based2021.12 | 72.9 | 83.6 | — | — | — | — | — | — | |
| MANGO2023.04 | 72.9 | 83.6 | — | — | — | — | — | — | |
| MANGOInput Size (longer side)=12802020.12 | 71.7 | 82.6 | 8.9 | — | — | — | — | — | |
| TESTR-BezierBackbone=ResNet-50, FPS=5.52022.04 | 71.6 | 83.3 | — | — | — | — | — | — | |
| Mask TextSpotter v3FPS=2.52021.05 | 71.2 | 78.4 | — | — | — | — | — | — | |
| Mask TextSpotter v3Multi-scale inference=false2022.03 | 71.2 | 78.4 | — | — | — | — | — | — | |
| Mask TextSpotter v3Backbone=ResNet-50-FPN2022.04 | 71.2 | 78.4 | — | — | — | — | — | — | |
| Mask TextSpotter v3Representation=Bounding Box-based2021.12 | 71.2 | 78.4 | — | — | — | — | — | — | |
| MaskTextSpotter v32023.04 | 71.2 | 78.4 | — | — | — | — | — | — | |
| Qin et al.Multi-scale inference=false2022.03 | 70.7 | — | — | — | — | — | — | — | |
| ABCNet v2FPS=102021.05 | 70.4 | 78.1 | — | — | — | — | — | — | |
| ABCNet-V22022.02 | 70.4 | 78.1 | — | — | — | — | — | — | |
| ABCNet v2Backbone=ResNet-50-FPN, FPS=102022.04 | 70.4 | 78.1 | — | — | — | — | — | — | |
| ABCNet v2Representation=Bounding Box-based2021.12 | 70.4 | 78.1 | — | — | — | — | — | — | |
| ABCNet v22023.04 | 70.4 | 78.1 | — | — | — | — | — | — | |
| ABCNet v2Paradigm=Supervised, Data Setting=Fully Labeled2025.04 | 70.4 | 78.1 | — | — | — | — | — | — | |
| Text PerceptronStage=End-to-end2020.02 | 69.7 | 78.3 | — | — | — | — | — | — | |
| Text Perceptron2020.12 | 69.7 | 78.3 | — | — | — | — | — | — | |
| Text PerceptronBackbone=ResNet-50-FPN2022.04 | 69.7 | 78.3 | — | — | — | — | — | — | |
| ABCNetData=SynText150k, COCO-Text, TT, MLT, Backbone=ResNet-50-FPN, Multi-scale testing=true2020.02 | 69.5 | 78.4 | 6.9 | — | — | — | — | — | |
| ABCNet-MSBackbone=ResNet-50-FPN, FPS=6.92022.04 | 69.5 | 78.4 | — | — | — | — | — | — | |
| CharNetMulti-scale inference=true2022.03 | 69.2 | — | — | — | — | — | — | — | |
| PAN++Multi-scale inference=false2022.03 | 68.6 | 78.6 | — | — | — | — | — | — | |
| PAN++Representation=Bounding Box-based2021.12 | 68.6 | 78.6 | — | — | — | — | — | — | |
| Qin et al.Data=SynText200k, IC15, COCO-Text, TT, MLT, Backbone=ResNet-50-MSF2020.02 | 67.8 | — | 4.8 | — | — | — | — | — | |
| Qin et al.FPS=4.82021.05 | 67.8 | — | — | — | — | — | — | — | |
| Unconstrained2020.12 | 67.8 | — | — | — | — | — | — | — | |
| UnconstrainedBackbone=ResNet-50-MSF2022.04 | 67.8 | — | — | — | — | — | — | — | |
| Qin et al.Representation=Bounding Box-based2021.12 | 67.8 | — | — | — | — | — | — | — | |
| ABCNet-V12022.02 | 67.2 | 76.4 | — | — | — | — | — | — | |
| CharNetRepresentation=Bounding Box-based2021.12 | 66.6 | — | — | — | — | — | — | — | |
| CharNetData=Private: 30k (manual label), 1m (partial label), Backbone=ResNet-50-Hourglass572020.02 | 66.2 | — | 1.2 | — | — | — | — | — | |
| CharNet R-50Backbone=ResNet-502020.12 | 66.2 | — | 1.2 | — | — | — | — | — | |
| CharNetBackbone=ResNet-50-Hourglass57, FPS=1.22022.04 | 66.2 | — | — | — | — | — | — | — | |
| Mask TextSpotter'19Data=SynText800k, IC13, IC15, TT, AddF2k, Backbone=ResNet-50-FPN2020.02 | 65.3 | 77.4 | 2 | — | — | — | — | — | |
| Mask TextSpotter’19FPS=22021.05 | 65.3 | 77.4 | — | — | — | — | — | — | |
| Mask TextSpotter2020.12 | 65.3 | 77.4 | 2 | — | — | — | — | — | |
| Mask TextSpotterRepresentation=Bounding Box-based2021.12 | 65.3 | 77.4 | — | — | — | — | — | — | |
| Boundary2020.12 | 65 | 76.1 | — | — | — | — | — | — | |
| Boundary TextSpotterBackbone=ResNet-50-FPN2022.04 | 65 | 76.1 | — | — | — | — | — | — | |
| ABCNetData=SynText150k, COCO-Text, TT, MLT, Backbone=ResNet-50-FPN2020.02 | 64.2 | 75.7 | 17.9 | — | — | — | — | — | |
| ABCNetFPS=17.92021.05 | 64.2 | 75.7 | — | — | — | — | — | — | |
| ABCNet2020.12 | 64.2 | 75.7 | 17.9 | — | — | — | — | — | |
| ABCNetRepresentation=Bounding Box-based2021.12 | 64.2 | 75.7 | — | — | — | — | — | — | |
| Text PerceptronStage=2-stage2020.02 | 63.3 | 73.9 | — | — | — | — | — | — | |
| PGNetBackbone=ResNet-50-FPN, FPS=35.52022.04 | 63.1 | — | — | — | — | — | — | — | |
| PGNetRepresentation=Bounding Box-based2021.12 | 63.1 | — | — | — | — | — | — | — | |
| ABCNetData=SynText800k, IC15, TT, Backbone=ResNet-50-FPN, Short side of input image=6002020.02 | 61.9 | 74.1 | 22.8 | — | — | — | — | — | |
| Li et al.Multi-scale inference=true2022.03 | 58.56 | — | — | — | — | — | — | — | |
| Li et al.Data=SynText840k, IC13, IC15, TT, MLT, AddF2k, Backbone=ResNet-101-FPN2020.02 | 57.8 | — | 1.4 | — | — | — | — | — | |
| Li et al.FPS=1.42021.05 | 57.8 | — | — | — | — | — | — | — | |
| Feng et al.FPS=7.22021.05 | 55.8 | 79.2 | — | — | — | — | — | — | |
| TextNetMulti-scale inference=false2022.03 | 54.02 | — | — | — | — | — | — | — | |
| TextNetData=SynText800k, IC13, IC15, TT, Backbone=ResNet-50-SAM2020.02 | 54 | — | 2.7 | — | — | — | — | — | |
| TextNetFPS=2.72021.05 | 54 | — | — | — | — | — | — | — | |
| TextNet2020.02 | 54 | — | — | — | — | — | — | — | |
| Mask TextSpotter' 18Data=SynText800k, IC13, IC15, TT, Backbone=ResNet-50-FPN2020.02 | 52.9 | 71.8 | 4.8 | — | — | — | — | — | |
| Mask TextSpotter’18FPS=2.62021.05 | 52.9 | 71.8 | — | — | — | — | — | — | |
| Mask TextSpotter2020.02 | 52.9 | 71.8 | — | — | — | — | — | — | |
| Mask TextSpotterMulti-scale inference=false2022.03 | 52.9 | 71.8 | — | — | — | — | — | — | |
| Mask TextSpotterBackbone=ResNet-50-FPN, FPS=4.82022.04 | 52.9 | 71.8 | — | — | — | — | — | — | |
| TextDragonData=SynText800k, IC15, MLT, TT, Backbone=VGG162020.02 | 48.8 | 74.8 | — | — | — | — | — | — | |
| TextDragonFPS=2.62021.05 | 48.8 | 74.8 | — | — | — | — | — | — | |
| TextDragon2020.12 | 48.8 | 74.8 | — | — | — | — | — | — | |
| TextDragonMulti-scale inference=false2022.03 | 48.8 | 74.8 | — | — | — | — | — | — | |
| Text DragonBackbone=VGG162022.04 | 48.8 | 74.8 | — | — | — | — | — | — | |
| Two-stageData=SynText800k, IC13, IC15, TT, Backbone=ResNet-50-SAM2020.02 | 45 | — | — | — | — | — | — | — | |
| TextBoxesData=SynText800k, IC13, IC15, TT, Backbone=ResNet-50-FPN2020.02 | 36.3 | 48.9 | 1.4 | — | — | — | — | — |