基于深度学习乳腺超声计算机辅助诊断系统研究进展  

Research progresses of breast ultrasound computer aided diagnosis systems based on deep learning

作　　者：吕明慧 周帅[1] 朱强[1] LYU Minghui;ZHOU Shuai;ZHU Qiang(Department of Diagnostic Ultrasound,Beijing Tongren Hospital,Capital Medical University,Beijing 100730,China)

机构地区：[1]首都医科大学附属北京同仁医院超声诊断科,北京100730

出　　处：《中国医学影像技术》2020年第11期1722-1725,共4页Chinese Journal of Medical Imaging Technology

基　　金：十三五国家重点研发计划项目(2016YFC0104803)。

摘　　要：超声是诊断及筛查乳腺癌的重要工具,为提高其诊断准确率,超声计算机辅助诊断(CAD)系统应运而生。传统的CAD系统需人工进行图像预处理及特征提取,工作量较大且诊断效能欠佳。深度学习(DL)利用计算机算法自动提取图像特征,较传统方法更接近人工智能,而其中应用较广的算法是卷积神经网络(CNN)。本文对乳腺CAD系统的发展及基于DL的乳腺超声CAD系统的研究进展进行综述。Ultrasonography is an important tool for diagnosis and screening of breast cancers.In order to improve the accuracy of diagnosis,ultrasound computer-aided diagnosis(CAD)system was established.The traditional version of CAD systems required image preprocessing and features extraction manually,needing to process a great deal of data and having lower efficiency of work.Deep learning(DL)makes the computer to automatically extract features and be more reflective of the essence of artificial intelligence than traditional methods.The most widely used algorithm is the convolutional neural network(CNN).The research progresses of breast ultrasound CAD systems based on DL were reviewed in this article.

关 键 词：乳腺肿瘤 图像处理 计算机辅助 深度学习 神经网络 计算机 超声检查 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] R737.9[自动化与计算机技术—控制科学与工程]