数字化乳腺超声图像数据库的初步建立与应用

摘要

Objective To create a large samples digital database of breast ultrasound imaging (DDBUI),which is open and can be used for scientific research,clinical training and diagnose in order to improve early diagnostic rate of breast cancer by using the database.Methods The database applies structured query language (SQL) to record the integrated resources of patients with breast masses and utilizes Brower/Server (B/S) technology to achieve universal data access.Results The database consists of 853 cases and 5780 images.So far,researchers in home and abroad have done many researches with the images in DDBUI.For example,the speckle reduction,enhancement,segmentation and classification algorithms for computer aided diagnosis (CAD).The database provides a complete shared operation platform for clinical diagnosis and teaching.Conclusions The first public database of breast ultrasound images is created which provides comprehensive information for the development and verification different CAD algorithms,and is helpful for researches,ultrasound diagnosis learning,and improving the breast cancer diagnosis.%目的 创建一个大样本、标准化的公共乳腺超声图像数据库,实现网上资源共享,为科研、临床及教学提供全面的数据图像信息,旨在提高乳腺癌的早期诊断率.方法 应用结构查询语言(structured query language,MySQL)数据库语言,将搜集的具有完整诊疗资料的乳腺肿块患者信息录入数据库,运用浏览器/服务器模式(Brower/Server,B/S)技术达成网络通用数据访问.结果 建立了包含853例5780张图像的乳腺超声图像数据库,应用图像特征信息研发出去噪、增强、分割及自动分类计算机辅助诊断算法,并为临床诊断、教学提供了完整的共享式操作平台.结论 建立了首个实现网上资源共享的乳腺超声图像数据库,为国内外研究机构研发与应用推广计算机辅助诊断(computer-aided diagnosis,CAD)算法提供标准统一的图像信息,辅助临床诊断和教学,对提高乳腺癌诊断率具有重要意义.