基于深度学习的帕金森病诊断  

Parkinson’s disease diagnosis based on deep learning

作　　者：谭言丹 赵阳洋 赵光财 TAN Yan-dan;ZHAO Yang-yang;ZHAO Guang-cai(Quanzhou Institute of Equipment Manufacturing,Haixi Institutes,Chinese Academy of Sciences,Quanzhou 362200,China;School of Electrical and Control Engineering,North University of China,Taiyuan 038507,China;School of Optoelectronic and Communication Engineering,Xiamen University of Technology,Xiamen 361024,China;School of Computer and Control Engineering,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院海西研究院泉州装备制造研究所,福建泉州362200 [2]中北大学电气与控制工程学院,山西太原038507 [3]厦门理工学院光电与通信工程学院,福建厦门361024 [4]中国科学院大学计算机与控制学院,北京100049

出　　处：《计算机工程与设计》2021年第8期2334-2340,共7页Computer Engineering and Design

摘　　要：为实现对帕金森病进行自动诊断,对基于Spiral和Meander手绘图的CNN和CapsNet的PD诊断方案进行研究。图像预处理过程保留了色彩信息,模型可以学习笔划压力、速度等特征。实验结果表明,CapsNet和CNN分别是基于Spiral和Meander的最优诊断方案。研究结果表明,Spiral具备比Meander更多的差异来诊断PD,CapsNet是基于Spiral的最佳方案;保留色彩信息的图像更有利于基于手绘图诊断PD,所实现结果(Acc=95.7%)优于先前最优报道(Acc=82.7%)。To achieve automatic diagnosis of PD,the PD diagnosis scheme of CNN and CapsNet based on Spiral and Meander hand-drawings was studied.The image preprocessing process retained color information so that the model could learn stroke pressure,speed,etc.The results show that the CapsNet and CNN are the optimal diagnosis schemes based on Spiral and Meander,respectively.The research shows two conclusions,the Spiral has more differences than the Meander to diagnose PD,and the CapsNet is the optimal solution based on Spiral.The image that retains the color information is more conducive to the PD diagnosis based on hand-drawing,and the achieved result(Acc=95.7%)is better than that of the previous best report(Acc=82.7%).

关 键 词：帕金森病 胶囊网络(CapsNet) 卷积神经网络(CNN) 深度学习 手绘图 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]