Authors: Yi Li, Weichen Zhang, Hui Song, Chunchao Guo, Ming Liu, Hongfa Wang, Lei Xiao

Description: Data Platform Department, Tencent. Our method is based on an ensemble of Faster-RCNN method and the traditional method (MSER). In the Faster-RCNN, we adapt a rotation-rpn which is helpful to detect inclined text region. The results of MESR method are used for adjusting the box confidence of the Faster-RCNN models.

Authors: Yi Li, Weichen Zhang, Hui Song, Chunchao Guo, Ming Liu, Hongfa Wang, Lei Xiao

Description: Data Platform Department, Tencent. Our method is based on Faster-RCNN architecture. In the Faster-RCNN, we adapt a rotation-rpn which is helpful to detect inclined text region.

Authors: Yuliang Liu, Lele Xie, Lianwen Jin, Sheng Zhang

Description: We used a two stages deep learning method to roughly recall text first and then finely adjust the bounding box for compact detection. A novel network was carefully designed for multi-scale training, and we used a learnable method to generate quadrilateral proposals for tightly matching the scene text. Finally, we designed a post-processing pipeline for further improving the precision.