Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ICDAR

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text DetectionICDAR 2015
Precision94.1
188
Scene Text DetectionICDAR 2015 (test)
F1 Score91.6
150
Oriented Text DetectionICDAR 2015 (test)
Precision94.1
129
Text DetectionICDAR 2015 (test)
F1 Score91.55
108
End-to-End Text SpottingICDAR 2015
Strong Score90.1
104
Text DetectionICDAR 2013 (test)
F1 Score92.5
88
End-to-End Text SpottingICDAR 2015 (test)
Generic F-measure81.9
62
Text DetectionICDAR Incidental Text 2015 (test)
Precision88.7
52
Scene Text RecognitionICDAR 2015 (test)
Accuracy87.24
46
Word SpottingICDAR 2015
Strong Score87.8
42
Scene Text RecognitionICDAR 2015
Accuracy (No Lexicon)93.1
41
Word SpottingICDAR 2015 (test)
F-score (Strong lexicon)88.8
36
Text SpottingICDAR 2015 (test)
Accuracy (Strong Lexicon)87.5
36
Scene Text RecognitionICDAR 2013
Accuracy97.8
33
Text LocalizationICDAR 2013 (test)
Recall87
28
Scene Text DetectionICDAR 2015
Precision92.5
25
End-to-End Text SpottingICDAR 2013 (test)
Score S95.94
25
Scene Text RecognitionICDAR13 (test)
Accuracy97.2
24
End-to-End RecognitionICDAR 2015 (test)
Strong Error Rate35
22
Scene Text RecognitionICDAR case-insensitive 2013 (test)
Accuracy93.3
22
Text RecognitionICDAR 2003
Accuracy98.7
19
Table DetectionICDAR 2013 (test)
Recall100
18
End-to-End recognitionICDAR 2015
Strong Error Rate (S)13.8
17
End-to-End Text SpottingICDAR 2003 (test)
F-measure91
17
Scene Text RecognitionICDAR 50-word lexicon 2003 (test)
Accuracy99.2
16
Showing 25 of 101 rows