点击上方“CVer”,选择加"星标"置顶
重磅干货,第一时间送达
本文转载自:CSIG文档图像分析与识别专委会
图4为LOTM结构。LOTM模块的输入Proposal Features是在Adaptive-RPN后的共享特征图上使用Deformable RoI pooling [4] 和双线性插值得到。经过1*1卷积后,展开两个平行分支,解耦为水平和和竖直两个正交方向的轮廓检测。水平方向分支使用1*k的卷积核水平方向卷积,竖直方向分支则使用k*1的卷积核竖直方向卷积,k是超参数,实验验证使用k=3比较好。卷积后的特征图经过Sigmoid归一化得到相应方向的热图。LOTM使用交叉熵损失分类轮廓边界点。
Point Re-scoring Algorithm模块中,先对两个方向热图进行简单的NMS预处理滤波得到更高置信度的准确表征,然后综合考虑LOTM输出的水平和垂直方向上响应,即文本轮廓需同时具有两个方向的响应,滤除单方向噪声,从而抑制伪召回。
三、主要实验结果及可视化效果Table 1. The single-scale results on Total-Text. * indicates the results from [5]. Ext is the short for external data used in training stage. y means testing at multi-scale setting. The evaluation protocol is DetEval.² ContourNet论文地址:https://arxiv.org/pdf/2004.04940.pdf
² ContourNet开源代码:https://github.com/wangyuxin87/ContourNet
参考文献[1] Tsung-Yi Lin, Piotr Doll´ar, Ross B. Girshick, Kaiming He, Bharath Hariharan, and Serge J. Belongie. Feature pyramid networks for object detection. In CVPR, pages 936–944, 2017.[2] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in neural information processing systems, pages 91–99, 2015.[3] Hamid Rezatofighi, Nathan Tsoi, JunYoung Gwak, Amir Sadeghian, Ian Reid, and Silvio Savarese. Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 658–666, 2019.[4] Xizhou Zhu, Han Hu, Stephen Lin, and Jifeng Dai. Deformable convnets v2: More deformable, better results. In CVPR, 2019.[5] Shangbang Long, Jiaqiang Ruan, Wenjie Zhang, Xin He, Wenhao Wu, and Cong Yao. Textsnake: A flexible representation for detecting text of arbitrary shapes. In ECCV, pages 19–35. Springer, 2018.[6] Yuliang Liu, Lianwen Jin, Shuaitao Zhang, Canjie Luo, Sheng Zhang.Curved scene text detection via transverse and longitudinal sequence connection. Pattern Recognition 90:337–345.[7] Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7132–7141, 2018.原文作者:Yuxin Wang, Hongtao Xie, Zhengjun Zha, Mengting Xing, Zilong Fu and Yongdong Zhang
撰稿:伍思航 | 编排:高 学审校:殷 飞 | 发布:金连文
免责声明:(1)本文仅代表撰稿者观点,撰稿者不一定是原文作者,其个人理解及总结不一定准确及全面,论文完整思想及论点应以原论文为准。(2)本文观点不代表本公众号立场。下载
在CVer公众号后台回复:CVPR2020,即可下载CVPR 2020所有论文和300+篇代码开源的论文项目,开源地址如下:
https://github.com/amusi/CVPR2020-Code
重磅!CVer-论文写作与投稿交流群成立
扫码添加CVer助手,可申请加入CVer-论文写作与投稿 微信交流群,目前已满2000+人,旨在交流顶会(CVPR/ICCV/ECCV/ICML/ICLR/AAAI等)、顶刊(IJCV/TPAMI等)、SCI、EI等写作与投稿事宜。
同时也可申请加入CVer大群和细分方向技术群,细分方向已涵盖:目标检测、图像分割、目标跟踪、人脸检测&识别、OCR、姿态估计、超分辨率、SLAM、医疗影像、Re-ID、GAN、NAS、深度估计、自动驾驶、强化学习、车道线检测、模型剪枝&压缩、去噪、去雾、去雨、风格迁移、遥感图像、行为识别、视频理解、图像融合、图像检索、论文投稿&交流、PyTorch和TensorFlow等群。
一定要备注:研究方向+地点+学校/公司+昵称(如论文写作+上海+上交+卡卡),根据格式备注,可更快被通过且邀请进群
▲长按加微信群
▲长按关注CVer公众号
整理不易,请给CVer一个在看!