朱东, 姜萍萍, 颜国正(), 王志武, 韩玎, 赵凯, 华芳芳, 姚盛健, 丁紫凡, 周泽润

上海交通大学 电子信息与电气工程学院, 上海 200240

收稿日期:2019-07-26出版日期:2020-08-28发布日期:2020-08-18
通讯作者:颜国正E-mail:gzhyan@sjtu.edu.cn
作者简介:朱 东(1993-),男,安徽省池州市人,硕士生,主要研究方向为人工肛门括约肌便意感知功能重建
基金资助:国家自然科学基金资助项目(61673271);国家自然科学基金资助项目(81601631);上海市科技支撑项目(19441910600);上海市科技支撑项目(19441913800)

Defecation Perception Reconstruction of an Artificial Anal Sphincter System

ZHU Dong, JIANG Pingping, YAN Guozheng(), WANG Zhiwu, HAN Ding, ZHAO Kai, HUA Fangfang, YAO Shengjian, DING Zifan, ZHOU Zerun

School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China

Received:2019-07-26Online:2020-08-28Published:2020-08-18
Contact:YAN Guozheng E-mail:gzhyan@sjtu.edu.cn

摘要/Abstract

摘要： 针对现有人造肛门括约肌系统在便意感知功能上数据获取方式和分析方法的缺陷,设计了一个便意感知重建系统.该系统主要包括传感器模块、数据采集存储模块的设计以及数据分析算法.通过多传感器与数据采集存储模块的配合重建患者直肠表面压力的分布变化情况,并提出基于主元分析(PCA)法和支持向量机(SVM)的便意预测模型(PCA-SVM).结果表明:上臂轴向和径向以及中臂径向位置的压力信号与便意产生有显著联系,选择基于高斯核函数的SVM算法,取惩罚因子C=0.0595 和核函数宽度σ=0.9536 对有效压力指标向量进行便意分类预测,与前馈神经网络模型相比,具有较高的预测准确度,满足人造肛门括约肌系统便意感知功能的要求.
关键词: 人工肛门括约肌, 便意感知重建, 主元分析, 支持向量机
Abstract: To overcome the shortcomings of data acquisition and analysis methods in the existing artificial anal sphincter system, a fecal perception reconstruction system is designed. The system mainly includes the design of a sensor module, a data acquisition and storage module, and a data analysis algorithm. Through the cooperation of the multi-sensor data acquisition and the storage module, the distribution of rectal surface pressure is reconstructed. A fecal perception prediction model based on principal components analysis (PCA) and support vector machines (SVM) is proposed (PCA-SVM). The results show that the axial and radial pressure signals of the upper arm and the radial position of the middle arm are significantly related to the fecal perception. The SVM algorithm based on the Gauss kernel function and the penalty factor C=0.0595 and the width σ=0.9536 pairs of effective pressure indicators are selected. Compared with the feed-forward neural network model, the vector model has a higher prediction accuracy and meets the requirements of the fecal perception function of the artificial anal sphincter system.
Key words: artificial anal sphincter, fecal perception reconstruction, principal component analysis, support vector machine (SVM)


PDF全文下载地址:

点我下载PDF