method: CLOVA-AI v22019-02-18
Authors: Jeonghun Baek, Junyeop Lee, Sungrae Park, Moonbin Yim, Seonghyeon Kim, Hwalsuk Lee
Description: We used Thin-plate-spline (TPS) based Spatial transformer network (STN) which normalizes the input text images, ResNet based feature extractor, BiLSTM, and attention mechanism.
This model was developed based on the analysis of scene text recognition modules.
See our paper and source code.
method: VARCO_v22020-12-15
Authors: Jusung Lee, Jaemyung Lee, Younghyun Lee, Joonsoo Lee
Affiliation: VARCO
Description: This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.1711097855, Text Localization and Recognition for Efficient Digital Contents Analysis)
method: Hancom Vision2020-09-04
Authors: Hancom Vision team
Description: Our model is featured by CNN-based, BiLSTM, and Attention.
Trained on MJSynthText + SynthText + external data (Pretrain), and Focused Scene Text 2013-2015.
Date | Method | Total Edit distance (case sensitive) | Correctly Recognised Words (case sensitive) | T.E.D. (case insensitive) | C.R.W. (case insensitive) | |||
---|---|---|---|---|---|---|---|---|
2019-02-18 | CLOVA-AI v2 | 26.9515 | 95.98% | 23.4836 | 96.35% | |||
2020-12-15 | VARCO_v2 | 36.8797 | 94.61% | 26.2131 | 95.89% | |||
2020-09-04 | Hancom Vision | 39.0466 | 93.52% | 33.6133 | 93.97% | |||
2018-09-12 | Clova AI / Lens | 39.1422 | 94.25% | 36.3422 | 94.61% | |||
2020-01-20 | VARCO | 39.8552 | 93.79% | 28.8885 | 95.07% | |||
2020-09-04 | Huawei_GDE_AI | 40.7449 | 90.32% | 32.5216 | 91.60% | |||
2020-09-30 | SRIB-STRIDE (Scene Text Recognition In Device) | 72.7929 | 86.67% | 62.6314 | 87.58% | |||
2016-01-25 | SRC-B-TextProcessingLab | 74.4549 | 87.40% | 63.1787 | 88.95% | |||
2015-12-29 | SRC-B-TextProcessingLab | 88.2727 | 84.66% | 78.2512 | 86.12% | |||
2015-11-09 | Megvii-Image++ | 115.9124 | 82.83% | 94.0676 | 86.03% | |||
2016-12-05 | SemaMediaData&HPI_Real-time-VideoOCR | 151.5163 | 82.37% | 123.2516 | 84.75% | |||
2016-11-29 | CNN-WRDF | 209.1936 | 72.66% | 193.4747 | 74.14% | |||
2013-04-08 | PicRead | 332.3699 | 57.99% | 290.8327 | 61.92% | |||
2013-04-05 | NESP | 355.2411 | 64.20% | 340.3159 | 64.84% | |||
2013-04-05 | PLT | 379.0842 | 62.37% | 362.2196 | 63.11% | |||
2013-04-05 | MAPS | 396.6110 | 62.74% | 380.8603 | 63.29% | |||
2013-04-09 | Feild's Method | 422.1187 | 47.95% | 390.6199 | 52.33% | |||
2013-04-06 | PIONEER | 479.8208 | 53.70% | 426.8353 | 55.71% | |||
2013-04-05 | TextSpotter | 606.2886 | 26.85% | 597.2748 | 28.13% |