List your company
ResearchFormerly WizWand

Scene Text Recognition on COCO-Text (test)

83AccuracyCLIP4STR-H
47.6456.826675.18Nov 2, 2018Aug 6, 2019May 9, 2020Feb 10, 2021Nov 14, 2021Aug 18, 2022May 23, 2023
Updated 8mo ago

Evaluation results

MethodLinks
CLIP4STR-HTrain data=RBU(6.5M)2023.0583
CLIP4STR-LTrain data=RBU(6.5M)2023.0582.7
CLIP4STR-LTrain data=Real(3.3M)2023.0581.9
CLIP4STR-BTrain data=RBU(6.5M)2023.0581.3
CLIP4STR-BTrain data=Real(3.3M)2023.0581.1
MPSTRTrain data=Real(3.3M)2023.0580.3
PARSeqTrain data=Real(3.3M)2023.0579.8
TRBATrain data=Real(3.3M)2023.0577.5
ABINetTrain data=Real(3.3M)2023.0576.5
DiG-ViT-BTrain data=Real(2.8M)2023.0575.8
ViTSTR-STrain data=Real(3.3M)2023.0573.6
Shi et al. 2018Lexicon=None, Training Level=word-level2018.1172.3
Yang et al. 2017Lexicon=None, Training Level=word-level and character-level2018.1169.3
TPS-ResNet-BiLSTM-AttnTrain Dataset=S90k+ST+TextOCR2021.0569.15
STAR-NetTrain Dataset=S90k+ST+TextOCR2021.0567.92
CLIP4STR-LTrain data=MJ+ST2023.0567
SARLexicon=None, Training Level=word-level2018.1166.8
CLIP-OCRTrain data=MJ+ST2023.0566.5
CLIP4STR-BTrain data=MJ+ST2023.0566.3
MPSTRTrain data=MJ+ST2023.0564.5
TPS-ResNet-BiLSTM-AttnTrain Dataset=TextOCR2021.0564.03
PARSeqTrain data=MJ+ST2023.0564
RosettaTrain Dataset=S90k+ST+TextOCR2021.0562.61
TRBATrain data=MJ+ST2023.0561.4
CRNNTrain Dataset=S90k+ST+TextOCR2021.0560.09
Liu, Chen, and Wong 2018Lexicon=None, Training Level=word-level and character-level2018.1159.3
ABINetTrain data=MJ+ST2023.0557.1
ViTSTR-STrain data=MJ+ST2023.0556.4
TPS-ResNet-BiLSTM-AttnPW=true, Train Dataset=S90k+ST2021.0556.34
STAR-NetPW=true, Train Dataset=S90k+ST2021.0553.56
TPS-ResNet-BiLSTM-AttnTrain Dataset=COCOText2021.0553.47
RosettaPW=true, Train Dataset=S90k+ST2021.0549.66
CRNNPW=true, Train Dataset=S90k+ST2021.0549