摘要
文本图像倾斜校正是文字识别前端的重要预处理环节。为了克服现有方法的倾斜角度检测范围只在-90°~90°的缺点,将文本图像倾斜角度检测问题转换为倾斜角度类检测问题,利用深度卷积神经网络的分类功能,选取适当的损失函数,设计了一阶段二分类和多阶段多分类的检测结构,实现了多种文本图像倾斜角度类的检测,进而得到了文本图像的倾斜角度范围。实验结果表明,倾斜角度类的准确率、召回率和精确率都在0.93以上。利用经典深度学习的方法对经过倾斜校正后的文本图像进行文字识别,识别精确率比校正前有大幅度的提升。
Text image skew correction is an important preprocessing step in the front-end of character recognition.To overcome the disadvantage in the limited range of tilt angle detection of the existing methods which is only-90--90°,this study transforms the text image skew angle detection problem into a skew angle class detection problem.Several types of skew angle classes of text images are detected using the classification function of deep convolution neural network by selecting the appropriate loss function and designing the detection structures of onestage two classification and multi-stage multi-classification,and then getting the tilt angle range of the text image.The experimental results show that the tilt angle class’s detection accuracy,recall,and precision rates are all above0.93.The classical deep learning method is used to recognize the text image after skew correction.Moreover,the recognition accuracy is greatly improved compared to that before the correction.
作者
郭从洲
李可
朱奕坤
童晓冲
王习文
Guo Congzhou;Li Ke;Zhu Yikun;Tong Xiaochong;Wang Xiwen(Department of Basic,Information Engineering UniversityZhengzhou,Henan 450001,China;School of Surveying and Mapping,Information Engineering University,Zhengzhou,Henan 450001,China)
出处
《激光与光电子学进展》
CSCD
北大核心
2021年第14期172-179,共8页
Laser & Optoelectronics Progress
基金
国家自然科学基金(41671409)。
关键词
图像处理
文本图像
倾斜角度
卷积神经网络
倾斜校正
文字识别
image processing
text image
skew angle
convolution neural network
skew correction
character recognition
作者简介
通信作者:郭从洲,czguo0618@sina.cn。
