method: Tencent-DPPR Team2019-04-30
Authors: Longhuang Wu, Shangxuan Tian, Chang Liu, Wenjie Cai, Jiachen Li, Sicong Liu, Haoxi Li, Chunchao Guo, Hongfa Wang, Hongkai Chen, Qinglin lu, Xucheng Yin, Lei Xiao
Description: Tencent-DPPR (Data Platform Precision Recommendation) Team. The method is based on two stage text detector, and use several different backbones and fcn based model to ensemble.
method: NJU_ImagineLab2019-05-01
Authors: Yao Xiao, Xiaoge Song, Wenhai Wang, Enze Xie, Tong Lu
Description: A instance segmentation-based method which is adopted from Mask R-CNN. Our model is trained on the joint dataset of LSVT, ICDAR2017-MLT and ICDAR2019-MLT.
ImagineLab, National Key Lab for Novel Software Technology, Nanjing University
Tongji University
method: SRCB_LSVT2019-04-29
Authors: Yi Yu, Haiyang Guo, Xiaobing Wang, Yingying Jiang
Description: We use Mask R-CNN for text detection in our submission, which extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition. Mask R-CNN is simple to train and adds only a small overhead to Faster R-CNN. And Mask R-CNN is the state-of-the-art object detection framework now. Therefore, it is used here for text detection task. We use the Mask R-CNN in https://github.com/facebookresearch/Detectron and the backbone network is ResNext 101. Meanwhile, the object is text here and the number of classification classes is 2. Besides, polygon based NMS is used for post-processing to remove overlapped text regions.
Date | Method | Recall | Precision | Hmean | |||
---|---|---|---|---|---|---|---|
2019-04-30 | Tencent-DPPR Team | 77.43% | 83.42% | 80.31% | |||
2019-05-01 | NJU_ImagineLab | 75.01% | 81.73% | 78.22% | |||
2019-04-29 | SRCB_LSVT | 72.39% | 83.10% | 77.38% | |||
2019-04-30 | A scene text detection method based on maskrcnn | 72.64% | 80.46% | 76.35% | |||
2019-04-30 | Fudan-Supremind Detection | 73.90% | 77.63% | 75.72% | |||
2019-04-30 | DMText_lsvt | 70.92% | 79.11% | 74.79% | |||
2019-04-26 | baseline_polygon_0.7 | 70.99% | 78.75% | 74.66% | |||
2019-04-30 | TMIS | 68.67% | 74.37% | 71.40% | |||
2019-04-30 | pursuer | 65.49% | 75.80% | 70.27% | |||
2019-04-30 | HUST_VLRGROUP | 69.17% | 71.38% | 70.26% | |||
2019-04-30 | PAT-S.Y | 63.08% | 74.31% | 68.24% | |||
2019-04-29 | VIC-LISAR | 61.47% | 74.26% | 67.27% | |||
2019-04-29 | test4 | 54.82% | 71.60% | 62.10% | |||
2019-04-26 | PSENet_v2 | 59.22% | 64.65% | 61.81% | |||
2019-04-26 | Papago OCR (PixelLink+) | 53.02% | 63.59% | 57.82% | |||
2020-11-23 | hrnet_w40_casecade | 56.84% | 57.33% | 57.08% | |||
2019-04-27 | JDIVA_Textboxes++ | 54.50% | 59.63% | 56.95% | |||
2019-04-20 | Simple Baseline | 54.92% | 56.14% | 55.52% | |||
2019-04-30 | CRAFT | 51.34% | 55.59% | 53.38% | |||
2022-03-03 | db | 45.19% | 59.56% | 51.39% | |||
2019-04-25 | AdvancedEast model with post processing | 43.20% | 52.52% | 47.41% | |||
2019-05-01 | TextMask_V1 | 68.16% | 28.82% | 40.51% | |||
2019-04-30 | DDT | 74.50% | 13.12% | 22.31% |