- Task 1 - Text Localization
- Task 2 - Script identification
- Task 3 - Joint text detection and script identification
- Task 4 - End-to-End text detection and recognition
method: SituTech_OCR2021-03-11
Authors: Kui Lyu, Chuanhe Liu
Affiliation: Beijing Situ Vision Technologies Co. Ltd
Email: lvkui@situdata.com
Description: In this work, we design an elegant text detection model. Our detector is similar to DBNet, but there are some difference. More specifically, we have introduced an advanced detector backbone, a classic network EfficientDet, with flexible scales and stronger ability to extract features. Another breakthrough is that we optimized the label generation strategy in DBNet. In the original work, the positive area generation and the expansion of the positive area to the bounding box used the Vatti clipping algorithm, which is less robust with different area perimeter ratios. We optimized this function to make the label transform between positive area and bounding box more reasonable.
If you have any questions, please contact us.
SituAIgorithm Team, Beijing Situ Vision Technologies Co. Ltd
method: Tencent-DPPR Team (Method_v0.1)2019-05-27
Authors: Longhuang Wu, Shangxuan Tian, Chang Liu, Wenjie Cai, Jiachen Li, Sicong Liu, Haoxi Li, Chunchao Guo, Hongfa Wang, Hongkai Chen, Qinglin lu, Chun Yang, Xucheng Yin, Lei Xiao
Description: We are Tencent-DPPR (Data Platform Precision Recommendation) team. Our method follows the framework of Mask R-CNN that employs mask to detect multi-oriented scene texts. We use the MLT-19 and the MSRA-TD500 dataset to train our text detector, and we also apply a multi-scale training approach during training. To obtain the final ensemble results, we combined two different backbones and different multi-scale testing approaches.
method: TH2020-04-19
Authors: Tsinghua University and Hyundai Motor Group AIRS Company
Email: Shanyu Xiao: xiaosy19@mails.tsinghua.edu.cn
Description: We have built an end-to-end scene text spotter based on Mask R-CNN & Transformer. The ResNeXt-101 backbone and multiscale training/testing are used.
| Date | Method | Hmean | Precision | Recall | Average Precision | |||
|---|---|---|---|---|---|---|---|---|
| 2021-03-11 | SituTech_OCR | 50.97% | 34.24% | 99.69% | 33.74% | |||
| 2019-05-27 | Tencent-DPPR Team (Method_v0.1) | 45.80% | 30.20% | 94.70% | 73.01% | |||
| 2020-04-19 | TH | 43.70% | 28.39% | 94.85% | 72.05% | |||
| 2019-11-11 | Sogou_OCR | 42.74% | 27.35% | 97.72% | 86.58% | |||
| 2023-05-22 | DeepSolo++ (ResNet-50) | 42.52% | 27.65% | 91.91% | 83.86% | |||
| 2024-05-30 | CPN (multi-scale) | 41.41% | 26.30% | 97.29% | 90.39% | |||
| 2019-05-27 | TH-DL | 41.36% | 26.81% | 90.37% | 50.59% | |||
| 2019-06-04 | multi-stage_text_detector_v4 | 40.22% | 25.54% | 94.62% | 25.42% | |||
| 2019-06-03 | TH-DL-v1 | 40.15% | 25.80% | 90.45% | 57.22% | |||
| 2019-06-03 | Tencent-DPPR Team (Method_v0.2) | 40.02% | 25.29% | 95.87% | 75.41% | |||
| 2019-06-04 | TH-DL-v2 | 39.87% | 25.58% | 90.37% | 56.76% | |||
| 2019-06-04 | Tencent-DPPR Team (Method_v0.3) | 39.64% | 24.99% | 95.83% | 75.00% | |||
| 2019-06-04 | Tencent-DPPR Team | 39.63% | 24.98% | 95.87% | 75.00% | |||
| 2019-06-03 | multi-stage_text_detector | 39.54% | 25.05% | 93.79% | 25.17% | |||
| 2019-06-04 | multi-stage_text_detector_v3 | 39.09% | 24.66% | 94.22% | 24.68% | |||
| 2019-06-04 | multi-stage_text_detector_v2 | 39.05% | 24.58% | 94.97% | 24.78% | |||
| 2019-06-03 | CRAFTS | 38.50% | 25.04% | 83.30% | 38.08% | |||
| 2022-11-02 | ESTextSpotter | 38.26% | 23.96% | 94.89% | 84.59% | |||
| 2019-05-30 | PMTD | 38.08% | 23.88% | 94.03% | 81.75% | |||
| 2019-06-03 | NJU-ImagineLab(v3) | 38.02% | 23.73% | 95.60% | 81.28% | |||
| 2019-06-03 | mm-maskrcnn_v2 | 37.85% | 23.73% | 93.52% | 84.12% | |||
| 2019-06-03 | TP | 37.63% | 24.31% | 83.22% | 23.93% | |||
| 2019-05-28 | Unicamp-SRBR-MLT2019-S1 | 37.42% | 24.71% | 77.09% | 41.96% | |||
| 2019-06-02 | DISTILLED CRAFT | 36.35% | 23.58% | 79.29% | 16.71% | |||
| 2019-05-31 | A two-stage text detector based on cascade rcnn | 34.10% | 21.02% | 90.22% | 77.70% | |||
| 2019-05-24 | PSENet_v1 | 34.04% | 21.56% | 80.79% | 30.96% | |||
| 2019-05-26 | two stage text detector | 33.39% | 20.49% | 90.18% | 75.77% | |||
| 2021-02-04 | NCU_MSP | 33.24% | 20.41% | 89.59% | 17.82% | |||
| 2019-06-02 | A two-stage text detector based on cascade rcnn(using total 10000 images of mlt19) | 32.95% | 20.13% | 90.73% | 77.59% | |||
| 2019-05-27 | CLTDR | 32.35% | 19.71% | 90.18% | 18.45% | |||
| 2020-10-16 | Drew | 31.92% | 19.86% | 81.34% | 62.29% | |||
| 2019-06-03 | sot | 31.20% | 18.69% | 94.54% | 81.58% | |||
| 2019-05-28 | CRAFTS(Initial) | 30.30% | 18.71% | 79.61% | 32.44% | |||
| 2019-06-04 | Unicamp-SRBR-MLT2019-PELEETEXT | 30.27% | 18.17% | 90.53% | 66.59% | |||
| 2019-05-29 | IC_RL | 29.65% | 17.70% | 91.39% | 16.90% | |||
| 2019-05-27 | MLT2019 ETD | 29.01% | 17.43% | 86.52% | 15.89% | |||
| 2019-05-29 | maskrcnn++ result | 28.98% | 17.15% | 93.40% | 16.92% | |||
| 2019-06-03 | RRPN | 28.78% | 17.01% | 93.20% | 74.54% | |||
| 2019-06-04 | Unicamp-SRBR-MLT2019-FUSION-PSENET-PELEETEXT | 28.18% | 16.67% | 91.20% | 63.64% | |||
| 2020-05-30 | NCU | 25.73% | 15.00% | 90.29% | 12.96% | |||
| 2019-06-03 | text-mountain | 25.09% | 14.51% | 92.73% | 14.20% | |||
| 2019-06-04 | Lomin OCR | 23.87% | 13.81% | 87.98% | 51.41% | |||
| 2019-05-27 | NXB OCR | 21.70% | 12.77% | 72.14% | 13.29% | |||
| 2020-10-07 | MEAST_V2_8_oct | 20.15% | 11.69% | 72.85% | 14.23% | |||
| 2023-05-30 | TD-PPIoU | 19.78% | 11.12% | 89.63% | 66.69% | |||
| 2020-10-23 | MEAST_V3_23_Oct | 18.64% | 10.70% | 72.34% | 12.75% | |||
| 2019-06-03 | NXB OCR | 17.97% | 10.08% | 82.40% | 8.47% | |||
| 2019-06-04 | Cyberspace | 17.01% | 9.98% | 57.29% | 5.50% | |||
| 2019-05-28 | PydBox-TextDetector | 9.52% | 6.88% | 15.44% | 1.14% | |||
| 2020-12-15 | DSIT-UOA | 3.95% | 2.08% | 38.19% | 1.73% | |||
| 2019-05-27 | Unicamp-SRBR-MLT2019-S1 | 0.02% | 0.02% | 0.04% | 0.00% | |||
| 2019-05-05 | AAAA | 0.01% | 0.01% | 0.04% | 0.00% | |||
| 2019-05-27 | 4Paradigm-Data-Intelligence | 0.00% | 0.00% | 0.00% | 0.00% | |||
| 2019-06-01 | tsinghuaee51_MLT2019 | 0.00% | 0.00% | 0.00% | 0.00% |