Robust ReadingCompetition
Challenges

method: Pixel-Anchor 2018-11-16

Authors: Yuan Li, Yuanjie Yu

Description: a novel end-to-end trainable deep neural network framework, named Pixel-Anchor, which combines semantic segmentation and SSD in one network by feature sharing and anchor-level attention mechanism to detect oriented scene text.

method: FOTS_v22018-01-22

Authors: Xuebo Liu, Ding Liang, Junjie Yan

Description: FOTS_v2 improves the performance on long text, such as Chinese.

method: SCUT_DLVClab12017-06-28

Authors: Yuliang Liu, Lianwen Jin, Sheng Zhang, Canjie Luo, Zhaohai Li, Lele Xie, Zenghui Sun

Description: Two deep learning-based stages have been used, the first is to roughly detect text, and the second is to finely adjust the bounding box for accurate detection. A novel network is designed for multi-scale learning to generate tight quadrilateral proposals for the scene text. This is followed by a post-processing pipeline to improve the precision.

Ranking Table

Description Paper Source Code
DateMethodAverage PrecisionPrecisionRecallHmean
2018-11-16Pixel-Anchor 27.27%31.97%70.19%43.93%
2018-01-22FOTS_v241.42%29.89%79.20%43.40%
2017-06-28SCUT_DLVClab152.71%29.86%71.92%42.20%
2018-10-29Amap-CVLab51.64%28.52%78.63%41.86%
2018-01-22FOTS40.07%29.11%73.07%41.64%
2017-11-09EAST++34.10%27.29%75.47%40.08%
2018-05-18PSENet_NJU_ImagineLab (single-scale)20.21%27.08%73.87%39.63%
2018-08-08PKU_VDIG53.06%26.90%74.04%39.46%
2018-08-23Sogou_MM57.38%23.34%78.67%36.00%
2018-03-12ATL Cangjie OCR35.50%20.84%69.98%32.11%
2017-06-29SARI_FDU_RRPN_v140.70%19.23%69.71%30.15%
2018-07-27tongjiu13.79%18.95%71.41%29.95%
2018-07-31east-IMPROVE19.63%15.75%63.92%25.27%
2017-06-28SARI_FDU_RRPN_v036.44%15.60%66.48%25.27%
2017-06-30TH-DL20.81%16.98%48.08%25.09%
2017-06-30linkage-ER-Flow12.37%10.30%54.72%17.34%
2017-06-30Sensetime OCR40.38%8.27%80.04%14.98%

Ranking Graphic

Ranking Graphic