method: SCUT-DLVC-Lab-Refinement2019-04-22

Authors: Jiapeng Wang*, Yan Li*, Tianwei Wang, Jiaxin Zhang, Yichao Huang, Canjie Luo, Kai Ding, Lianwen Jin (*equal contribution), SCUT-DLVC-Lab & INTSIG-AIM

Description: This method utilizes a refinement-based Mask-RCNN. We iteratively remove redundant information and refine detected bounding-box results. Finally, the submitted result is ensembled from several models with different backbones.

Authors: Xianbiao Qi, Ning Lu, Yuan Gao, Yihao Chen, Shaoqiong Chen, Wenwen Yu, Rong Xiao

Description: We use an anchor-free detection framework with FishNet as our backbone. We pre-process the training and testing images to roughly align their scales with OpenCV adaptive threshold. We post-process the detection results according to the classification score map. It can well solve the problem of short detection of long text line.

1. Sun, Shuyang, et al. "Fishnet: A versatile backbone for image, region, and pixel level prediction." Advances in Neural Information Processing Systems. 2018. 2. Zhou, Xinyu, et al. "EAST: an efficient and accurate scene text detector." Proceedings of the IEEE conference on Computer Vision and Pattern Recognition. 2017. 3. Zhang, Zhi, et al. "Bag of Freebies for Training Object Detection Neural Networks." arXiv preprint arXiv:1902.04103 (2019).

method: H&H Lab2019-04-22

Authors: HUST_VLRGROUP(Mengde Xu, Zhen Zhu,Hui Zhang, Mingkun Yang, Jiehua Yang) & HUAWEI_CLOUD_EI(Jing Wang, Yibin Ye, Shenggao Zhu, Dandan Tu)

Description: we ensemble EAST and multi oriented corner to create a robust scene text detector. To make network learning easier, we modified the mutli-oriented corner network with a new branch borrowed from east added.

Ranking Table

Description Paper Source Code
2019-04-22Ping An Property & Casualty Insurance Company98.60%98.40%98.50%
2019-04-22H&H Lab97.93%97.95%97.94%
2019-04-22GREAT-OCR Shanghai University96.62%96.21%96.42%
2019-05-10Clova OCR96.04%95.79%95.92%
2019-04-22A Single-Shot Model for Robust Text Localization93.93%94.80%94.37%
2019-04-23SROIE Fourth Submission92.98%94.99%93.97%
2019-04-22Pixellink multi-scale Detection93.07%92.84%92.95%
2019-04-19BiLSTM Based on CTPN91.40%94.03%92.69%
2019-04-22CITlab Argus Textline Detection92.02%91.34%91.68%
2019-04-19Unet and Morphology Prediction93.28%89.43%91.31%
2019-04-17Textline detection89.85%92.72%91.26%
2019-04-20A Text Localization Method Based on CTPN85.23%88.73%86.94%
2019-04-22YOLO Text Detector77.29%79.32%78.29%
2019-04-17Improved yolov3 model68.52%78.23%73.06%
2019-04-22Task 1 - Scanned Receipt Text Localisation (Submitted by Intuit Inc.)71.14%63.76%67.25%
2019-04-17scene text detection weapon49.61%64.75%56.18%
2019-04-22Unet Segmentation and Watershed56.31%53.46%54.85%
2019-04-22Receipt Info Extracting Task1 zone-dividing32.62%46.48%38.33%

Ranking Graphic