Authors: 杨超杰(Chaojie Yang) 柳博(Bo Liu) 郑伟(Wei Zheng) 张明月(Mingyue Zhang) 刘姝辰(Shuchen Liu) 崔炜炜(Weiwei Cui) 苏统华(TongHua Su)

Description: Our team is from Harbin Institute of Technology.Contact as below:15754605370@163.com.My methodolopy use textboxes++ structure basiclly,which detects arbitrary-oriented scene text with both high accuracy and efficiency in a single network forward pass.It inherits the popular VGG-16 architecture , keeping the layers from conv1_1 through conv5_3, and converting the last two fully-connected layers of VGG-16 into convolutional layers (conv6 and conv7) by parameters down-sampling.