摘要
为实现笔画的分组和识别,现有的草图识别算法通常会采用限制用户的绘图习惯来达到目的.该文提出了利用贝叶斯网络和卷积神经网络(CNN)的草图识别方法解决此问题.首先,使用高斯低通滤波器处理输入草图,得到更平滑的图像.然后将连续输入的笔划分为两部分,分别使用贝叶斯网络和卷积神经网络对分割后的笔画进行识别,当笔画的可靠性大于阈值时,以贝叶斯网络的识别结果为准,反之采用CNN的识别结果.实验结果表明,该文算法在电路符号绘制过程中的识别率和绘制完成后的识别率均取得了较好的结果.该文算法具有良好的应用前景.
Most of the existing sketch recognition algorithms have been used to restrict the user s drawing habits to achieve the stroke grouping and recognition. In this paper, a sketch recognition method based on Bayesian network and convolutional neural network (CNN) is proposed to solve this problem. First, the input sketch is processed by gaussian low-pass lter and a smoother stroke can be obtained. The stroke of continuous input is divided, then the bayesian network and CNN are performed on stroke recognition respectively. The recognition result of bayesian network is adopted when the reliability of stroke is larger than the threshold, otherwise recognition result of CNN will be adopted. The experiment result shows that the proposed algorithm is effective in circuit symbol recognition. The recognition rate was achieved 81.27% in the drawing process, and the final recognition rate was achieved 92.16%.
作者
李鸿雁
苏庭波
LI Hong-yan;SU Ting-bo(Shangqiu university applied science and technology college,Shangqiu Henan 476000, China;School of Computer Engineering, Shangqiu University, Shangqiu Henan 476000, China)
出处
《西南师范大学学报（自然科学版）》
CAS
北大核心
2019年第9期96-102,共7页
Journal of Southwest China Normal University(Natural Science Edition)
基金
河南省科技厅项目(182102210511)
关键词
贝叶斯网络
卷积神经网络
笔画分组
草图识别
Bayesian network
convolution neural network
stroke grouping
sketch recognition
