Sample RankingRecallPrecisionHmean
Sensetime V1100.00%100.00%100.00%
FOTS100.00%100.00%100.00%
Baidu VIS v2100.00%100.00%100.00%
Tencent Youtu100.00%100.00%100.00%
CPN (multi-scale)100.00%100.00%100.00%
Baidu IDL v388.89%100.00%94.12%
Baidu VIS88.89%100.00%94.12%
SRC-B-MachineLearningLab88.89%100.00%94.12%
Baidu VIS v2+88.89%100.00%94.12%
Sogou_MM88.89%100.00%94.12%
gts-swint-cocoonly-s4-3_188.89%100.00%94.12%
Alibaba-PAI88.89%88.89%88.89%
HUST_MCLAB66.67%100.00%80.00%
HoText_v177.78%77.78%77.78%
e2e-mlt + bug fix77.78%70.00%73.68%
SRC-B-MachineLearningLab66.67%75.00%70.59%
Baidu IDL66.67%66.67%66.67%
Smile55.56%83.33%66.67%
NVTextSpotter66.67%60.00%63.16%
TextProposals + DictNet44.44%100.00%61.54%
SRC-B-TextProcessingLab33.33%100.00%50.00%
DIY FOTS44.44%36.36%40.00%
Stradvision-222.22%66.67%33.33%
Yunos_Robot1.022.22%66.67%33.33%
M_FOTS_Orig33.33%33.33%33.33%
StradVision_v133.33%30.00%31.58%
e2e-mlt33.33%30.00%31.58%
Char-level PTSD on SynthText22.22%40.00%28.57%
Megvii-Image++22.22%22.22%22.22%
Baseline OpenCV3.0 + Tesseract11.11%50.00%18.18%
NJU Text (Version3)11.11%20.00%14.29%
Baseline-TextSpotter0.00%0.00%0.00%
Beam search CUNI +S (decoding for TextSpotter - no postprocessing)0.00%0.00%0.00%
Beam search CUNI (decoding for TextSpotter - spell checked)0.00%0.00%0.00%
Deep2Text-MO0.00%0.00%0.00%