method: HIT2020-05-13
Authors: Sihwan Kim and Taejang Park
Affiliation: Hana Institute of Technology
Description: we present the network architecture to maximize conditional log-likelihood by optimizing the lower bound with a proper approximate posterior that has shown impressive performance in several generative model. In addition, by extending layer of latent variables to multiple layers, the network is able to learn scale robust features with no task specific regularization or data augmentation. We provide a detailed analysis and show the results of three public benchmarks to confirm the efficiency and reliability of the proposed algorithm.
method: Craft++2020-05-19
Authors: Xiangyuan Ren, Anjie Song, Zikun Zhou
Affiliation: Shanghai Jiao Tong University, ShannonAi
Email: xiangyuan_ren@shannonai.com
Description: Out Method is based on CRAFT, with Self Supervised Learning for pretraining and stroke level segmentation for multi-task training
method: VARCO2020-12-15
Authors: Jaemyung Lee, Jusung Lee, Younghyun Lee, Joonsoo Lee
Affiliation: NCSOFT
Description: This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.1711117050, Text Localization and Recognition for Efficient Digital Contents Analysis)
Description Paper Source Code
| Date | Method | Recall | Precision | Hmean | |||
|---|---|---|---|---|---|---|---|
| 2020-05-13 | HIT | 94.36% | 98.48% | 96.38% | |||
| 2020-05-19 | Craft++ | 94.67% | 96.44% | 95.54% | |||
| 2020-12-15 | VARCO | 92.62% | 97.54% | 95.02% | |||
| 2018-11-07 | CRAFT | 92.40% | 97.67% | 94.96% | |||
| 2020-07-31 | TextFuseNet | 92.09% | 97.27% | 94.61% | |||
| 2020-01-20 | VARCO | 92.49% | 94.74% | 93.60% | |||
| 2018-04-17 | Ali-Amap-xlab-v4 | 91.67% | 94.99% | 93.30% | |||
| 2021-01-21 | NCU_MSP | 90.79% | 95.83% | 93.24% | |||
| 2017-11-21 | TencentAILab | 94.65% | 91.87% | 93.24% | |||
| 2020-01-21 | ChinaUnicom-AI | 91.34% | 94.46% | 92.87% | |||
| 2019-07-23 | CM-CV&AR | 91.42% | 93.98% | 92.68% | |||
| 2020-11-10 | Hancom Vision | 88.27% | 97.24% | 92.54% | |||
| 2018-01-22 | FOTS | 90.47% | 94.63% | 92.50% | |||
| 2018-12-03 | SPCNet_TongJi & UESTC (single scale) | 90.59% | 93.77% | 92.16% | |||
| 2020-01-06 | NCU_MSP | 91.63% | 92.62% | 92.12% | |||
| 2017-08-10 | SRC-B-MachineLearningLab-v4 | 89.77% | 93.73% | 91.71% | |||
| 2019-07-12 | stela | 89.66% | 93.74% | 91.65% | |||
| 2017-06-25 | FEN | 89.26% | 94.09% | 91.61% | |||
| 2019-05-29 | crpn.v2 | 89.55% | 93.59% | 91.53% | |||
| 2019-05-03 | Mask Textspotter | 88.27% | 95.01% | 91.52% | |||
| 2017-02-17 | NLPR-CASIA | 88.55% | 94.59% | 91.47% | |||
| 2016-12-16 | RRPN-4 | 87.85% | 94.91% | 91.25% | |||
| 2017-12-15 | EPTN-SJTU | 89.02% | 93.17% | 91.05% | |||
| 2019-02-12 | NCSOFT VISION AI LAB | 91.73% | 90.09% | 90.90% | |||
| 2018-03-05 | HappyCCL | 88.13% | 92.09% | 90.06% | |||
| 2017-03-16 | Ali-Amap-xlab-v2 | 87.40% | 91.80% | 89.54% | |||
| 2022-08-26 | HBLAB-OCR | 85.94% | 93.19% | 89.42% | |||
| 2018-05-06 | cvmt_allStepNoLkTree | 88.64% | 90.16% | 89.39% | |||
| 2017-03-23 | RTN | 85.08% | 93.85% | 89.25% | |||
| 2018-05-06 | CVMT_frm62 | 87.67% | 89.73% | 88.69% | |||
| 2016-12-04 | Ali-Amap-xlab | 86.16% | 90.89% | 88.46% | |||
| 2016-06-23 | Baidu IDL | 84.51% | 92.68% | 88.41% | |||
| 2017-11-15 | Sensetime line-level detection | 81.52% | 96.48% | 88.37% | |||
| 2017-01-22 | SRC-B-MachineLearningLab | 83.20% | 93.06% | 87.85% | |||
| 2017-03-22 | MCLAB_TextBoxes_v2 | 84.38% | 91.21% | 87.67% | |||
| 2018-01-04 | crpn | 83.80% | 91.90% | 87.66% | |||
| 2017-07-04 | FPTD | 86.94% | 88.35% | 87.64% | |||
| 2021-01-09 | FENET | 79.67% | 97.09% | 87.52% | |||
| 2017-05-02 | TsinghuaOCR | 84.37% | 89.75% | 86.98% | |||
| 2016-01-27 | SenseTime | 85.22% | 88.77% | 86.96% | |||
| 2019-03-23 | SSP-RPNs with Pytorch | 87.73% | 85.51% | 86.60% | |||
| 2016-07-18 | SCUT-HCII | 82.59% | 90.40% | 86.32% | |||
| 2017-04-04 | xmu403 | 82.83% | 89.95% | 86.24% | |||
| 2017-02-21 | STDN-2 | 78.81% | 95.07% | 86.18% | |||
| 2016-12-21 | MSRA_v1 | 79.87% | 93.32% | 86.08% | |||
| 2017-02-28 | Tencent Youtu | 79.47% | 93.87% | 86.07% | |||
| 2018-12-08 | Unicamp-SRBR-v2 | 80.97% | 91.80% | 86.05% | |||
| 2019-04-19 | CenterText(Single-scale) | 83.16% | 88.75% | 85.86% | |||
| 2017-02-07 | CAS_HotEye | 78.63% | 93.69% | 85.50% | |||
| 2017-02-27 | CTDN-3-PN | 77.46% | 94.85% | 85.28% | |||
| 2017-08-15 | MultDet | 83.80% | 86.42% | 85.09% | |||
| 2016-08-31 | MCLAB_TextBoxes | 82.59% | 87.73% | 85.08% | |||
| 2016-06-23 | SRC-B-TextProcessingLab | 80.18% | 90.50% | 85.03% | |||
| 2016-04-10 | SCUT-HCII | 82.79% | 87.17% | 84.93% | |||
| 2017-02-25 | CTDN2 | 76.11% | 95.20% | 84.59% | |||
| 2018-05-06 | cvmt_frm59 | 86.45% | 82.19% | 84.27% | |||
| 2015-04-03 | StradVision | 78.85% | 90.21% | 84.15% | |||
| 2017-05-27 | BUPT_NIRC_multi | 79.07% | 89.34% | 83.89% | |||
| 2016-03-16 | TextConv+WordGraph | 75.85% | 93.30% | 83.68% | |||
| 2017-05-27 | BUPT_NIRC_multi | 77.94% | 90.23% | 83.63% | |||
| 2015-04-02 | VGGMaxNet_025 | 79.58% | 88.04% | 83.59% | |||
| 2015-03-26 | VGGMaxNet_cmb | 77.57% | 90.47% | 83.53% | |||
| 2015-11-04 | MSER_Binary_CNN | 78.67% | 88.79% | 83.42% | |||
| 2016-03-16 | TextConv+WordGraph | 75.23% | 93.28% | 83.29% | |||
| 2015-03-23 | VGGMaxNet_013 | 76.49% | 91.36% | 83.27% | |||
| 2019-01-31 | ssprpns | 76.38% | 91.26% | 83.16% | |||
| 2015-11-28 | CASIA_USTB-Cascaded | 77.63% | 89.50% | 83.14% | |||
| 2017-03-05 | WeText | 76.40% | 91.16% | 83.13% | |||
| 2015-03-23 | VGGMaxNet_1.6 | 75.71% | 91.85% | 83.00% | |||
| 2017-03-24 | (๑•̀ㅂ•́)و✧ | 82.34% | 83.54% | 82.94% | |||
| 2015-10-21 | MCLAB_FCN | 77.81% | 88.14% | 82.65% | |||
| 2018-12-08 | Unicamp-SRBR-v3 | 74.92% | 91.38% | 82.34% | |||
| 2016-11-08 | CTPN | 73.72% | 92.77% | 82.15% | |||
| 2017-03-01 | STDN | 78.90% | 85.64% | 82.14% | |||
| 2016-01-18 | Text-CNN | 72.89% | 92.79% | 81.65% | |||
| 2019-07-17 | AFCTPN | 71.82% | 94.55% | 81.63% | |||
| 2020-03-26 | RRPN R-50 model_final 20200327 | 77.17% | 86.30% | 81.48% | |||
| 2017-04-01 | ConnLink | 74.16% | 90.14% | 81.37% | |||
| 2017-04-06 | ConnLink_pre | 71.98% | 92.26% | 80.87% | |||
| 2017-07-24 | Jack's TD | 73.99% | 89.06% | 80.83% | |||
| 2017-05-29 | ssd pretrain on synthtext and scut | 79.51% | 81.85% | 80.66% | |||
| 2019-06-26 | std(single-scale) | 77.13% | 84.48% | 80.64% | |||
| 2017-05-29 | FCN based network for Text Detection | 74.94% | 87.16% | 80.59% | |||
| 2017-09-05 | P-SSD.v1 | 80.47% | 80.68% | 80.58% | |||
| 2014-11-12 | HUST_MCLAB | 74.28% | 87.68% | 80.43% | |||
| 2016-11-13 | RRPN-3 | 71.89% | 90.22% | 80.02% | |||
| 2015-01-01 | BUCT_YST | 72.84% | 84.43% | 78.21% | |||
| 2017-06-09 | MPT+Jar | 68.16% | 89.25% | 77.29% | |||
| 2018-07-03 | ACI | 70.59% | 84.76% | 77.03% | |||
| 2014-06-10 | IWRR2014 | 70.01% | 85.61% | 77.03% | |||
| 2014-05-13 | SWT | 73.15% | 80.83% | 76.80% | |||
| 2017-08-25 | bupt | 67.96% | 88.24% | 76.79% | |||
| 2018-07-17 | 0.956_600_1000 | 69.30% | 85.73% | 76.64% | |||
| 2018-07-17 | 0.956_1280_960 | 67.27% | 88.88% | 76.58% | |||
| 2018-07-17 | 123_0.927_1280_960 | 70.72% | 82.87% | 76.31% | |||
| 2013-04-07 | USTB_TexStar | 66.45% | 88.47% | 75.89% | |||
| 2017-04-26 | Cascade Filtering and Grouping | 69.06% | 82.96% | 75.38% | |||
| 2019-02-16 | [BKU K15] | 73.24% | 77.26% | 75.20% | |||
| 2013-04-05 | TextSpotter | 64.84% | 87.51% | 74.49% | |||
| 2022-05-08 | 11111 | 77.86% | 71.23% | 74.40% | |||
| 2013-08-29 | UMD_IntegratedDisrimination | 62.26% | 89.17% | 73.33% | |||
| 2013-04-08 | CASIA_NLPR | 68.24% | 78.89% | 73.18% | |||
| 2018-12-08 | Unicamp-SRBR-v1 | 62.56% | 86.81% | 72.71% | |||
| 2015-07-22 | ZText | 63.51% | 84.73% | 72.60% | |||
| 2013-12-23 | BayesText | 63.51% | 84.30% | 72.44% | |||
| 2013-04-08 | Text_detector_CASIA | 62.85% | 84.70% | 72.16% | |||
| 2013-04-09 | I2R_NUS_FAR | 69.00% | 75.08% | 71.91% | |||
| 2017-10-12 | TextFCN V2 | 75.53% | 67.75% | 71.43% | |||
| 2017-08-13 | P-SSD.v1 | 65.52% | 77.51% | 71.01% | |||
| 2014-08-18 | DetectText | 66.05% | 75.82% | 70.60% | |||
| 2015-03-23 | VGGMaxNet_10 | 54.87% | 97.35% | 70.18% | |||
| 2013-04-08 | I2R_NUS | 66.17% | 72.54% | 69.21% | |||
| 2013-04-08 | TH-TextLoc | 65.19% | 69.96% | 67.49% | |||
| 2018-12-18 | test | 58.01% | 75.63% | 65.66% | |||
| 2018-12-29 | fast_ret_sh_02 | 55.32% | 76.09% | 64.06% | |||
| 2017-04-02 | CNN_Text | 65.68% | 59.93% | 62.67% | |||
| 2015-12-17 | TEST17 | 52.46% | 77.05% | 62.42% | |||
| 2013-04-06 | Text Detection | 53.42% | 74.15% | 62.10% | |||
| 2017-03-24 | CNN | 64.91% | 57.86% | 61.18% | |||
| 2015-08-18 | MSER with LocalSWT | 45.37% | 65.05% | 53.46% | |||
| 2014-09-16 | MSERs | 51.78% | 52.48% | 52.13% | |||
| 2015-08-18 | MSER | 44.62% | 53.31% | 48.58% | |||
| 2013-04-25 | Baseline | 34.74% | 60.76% | 44.21% | |||
| 2013-04-10 | Inkam | 35.27% | 31.20% | 33.11% |