method: Alibaba-PAI2019-02-01

Authors: Alibaba-PAI OCR team

Description: A end-to-end text spotter by Alibaba-PAI OCR team.

method: Sogou_MM2018-11-02

Authors: Baojun Li, Tao Wei, Qian Wang, Bin Li

Description: An End-to-End Text Spotter with Global Concatenating Feature Enhancement.

method: Baidu VIS v2+2018-08-01

Authors: VIS-OCR

Description: An improved end-to-end text localization and recognition result submitted by Baidu VIS-OCR.

Ranking Table

Description Paper Source Code
DateMethodRecallPrecisionHmean
2019-02-01Alibaba-PAI70.15%77.50%73.64%
2018-11-02Sogou_MM66.06%78.58%71.78%
2018-08-01Baidu VIS v2+66.59%75.33%70.69%
2018-07-30Tencent Youtu60.42%79.83%68.79%
2018-07-02Baidu VIS v266.59%69.95%68.23%
2018-03-28SRC-B-MachineLearningLab61.72%74.62%67.56%
2017-10-23Baidu VIS63.02%71.37%66.94%
2018-03-05HoText_v163.46%68.36%65.82%
2018-01-22FOTS53.20%84.61%65.33%
2017-08-02Sensetime V160.57%68.74%64.40%
2017-07-17Baidu IDL v360.23%66.16%63.05%
2018-07-25Smile57.05%58.09%57.57%
2019-04-11e2e-mlt + bug fix55.71%58.49%57.07%
2017-07-12SRC-B-MachineLearningLab50.99%59.63%54.97%
2018-10-25e2e-mlt49.45%55.13%52.13%
2016-07-21Baidu IDL48.87%52.70%50.71%
2016-10-20Yunos_Robot1.036.30%67.81%47.29%
2016-03-02TextProposals + DictNet33.61%79.14%47.18%
2016-04-05SRC-B-TextProcessingLab28.36%48.52%35.79%
2021-02-11M_FOTS_Orig32.50%35.81%34.07%
2015-11-09Megvii-Image++27.68%40.41%32.86%
2021-05-25Char-level PTSD on SynthText18.73%33.19%23.95%
2015-04-02Beam search CUNI +S (decoding for TextSpotter - no postprocessing)12.37%29.64%17.46%
2015-04-02Deep2Text-MO13.82%21.34%16.77%
2015-05-04Baseline OpenCV3.0 + Tesseract5.06%19.30%8.01%
2015-04-02Beam search CUNI (decoding for TextSpotter - spell checked)3.80%34.96%6.86%
2015-04-02Baseline-TextSpotter0.00%0.00%0.00%

Ranking Graphic