董建民 刘建新
摘 要:针对计算机辅助诊断中医学影像纹理计算相对复杂、商业分析软件功能相对固定和成本较高等问题,以目前主流的Python语言为开发工具,将其对应的PyRadiomics模块、SimpleITK模块与Pydicom模块相结合,给出了符合Dicom 3.0格式的医疗影像数据纹理分析软件设计流程,开发了纹理计算过程和数据可视化原型系统。系统集成为影像医师诊断提供了一种低成本、交互友好的纹理特征数据获取与分析手段,为医学影像中的人工智能、深度学习应用打下了基础。
关键词:PyRadiomics;医疗影像;纹理获取
DOI:10. 11907/rjdk. 201074 开放科学(资源服务)标识码(OSID):
中图分类号:TP317.4 文献标识码:A 文章编号:1672-7800(2020)007-0223-04
Integration of Medical Image Texture Acquisition Prototype System
Based on PyRadiomics
DONG Jian-min,LIU Jian-xin
(School of Information Engineering,Xizang Minzu University, Xianyang 712082,China)
Abstract:In order to solve the problems of relatively complex texture calculation, relatively fixed function and high cost of commercial analysis software in computer-aided diagnosis through medical image,after the design flow of texture analysis of medical image data been given, a prototype system of texture calculation is integrated by Python with its corresponding pyradiomics module, simpleITK module and pydicom module. The system could process all DICOM (digital imaging and communications in medicine) 3.0 format files of medicine images and texture calculation and visualization model are developed and a low-cost, interactive and friendly method of texture feature data display and analysis for imaging doctors diagnosis is provided. It lays a solid foundation for the fields of artificial intelligence and deep learning based on medical image.
Key Words:PyRadiomics;medical imaging;texture acquisition
0 引言
随着医疗设备数字化的不断深入,从数据管理技术上看,以影像归档和通信系统(Picture Archiving and Communication Systems,PACS)为依托,将计算机断层(Computed Tomography,CT)、核磁共振(Magnetic Resonance Imaging,MRI)、超声(Ultrasound,US)成像等各种医疗影像设备形成以Dicom(Digital Imaging and Communications in Medicine) 3.0为标准的医疗影像数据标准,目前已经基本实现了统一的数据交换和管理模式[1]。从已有医疗影像采集设备看,图像质量从空间上越来越精细,采集所需时间越来越短,图像重建速度也越来越快,并且整个过程对被检查对象几乎没有任何损害,这为医疗影像应用从