一种改进的电子病历系统的设计与开发

Design & Development of an Improved Electronic Medical Records System

针对以往电子病历系统中临床辅助诊疗功能存在的不足,系统做了两方面改进:首先是医疗影像模块中加入了人机交互,提出了一种基于AdaBoost-SVM算法的相关反馈机制,解决了CBMIR中上层语义与底层特征之间的语义鸿沟问题,同时提高了相关反馈中机器自主学习的能力,减轻了医生进行图像标注的负担,通过实验证明了该算法的有效性。其次是医疗数据的集成视图展现,将病人的基本信息、就诊信息以时间轴的形式展现,设计了一套可视化组件生成个性化数据集成,解决了医疗信息过载的问题。

This paper gives two improvements：firstly,medical imaging browser has enhanced the ability of human-computer inter- action,by adding Adaboost-SVM algorithm-based relevance feedback mechanism to solve the semantic gap between the upper semantics and underlying characteristics in CBMIR,and improve the machine independent learning ability in relevance feedback;Secondly,this system has joined medical data integrated view of medical data,Patient information and treatment information has been shown in one interface.We can ultimately form a set of integrated data views,and provide a feasible solution to prevent medical information overload.