基于神经网络的肝脏图像特征提取应用研究  

Application Research on Feature Extraction of Liver Image Based on Neural Network

作　　者：汪辉进[1] 曾叶纯[2] 徐宝娣 WANG Huijin;ZENG Yechun;XU Baodi(Department of Electronic Information and Media,Chizhou Vocational and Technical College,Chizhou Anhui 247000,China;Department of Biology and Health,Chizhou Vocational and Technical College,Chizhou Anhui 247000,China)

机构地区：[1]池州职业技术学院电子信息与传媒系,安徽池州247000 [2]池州职业技术学院生物与健康系,安徽池州247000

出　　处：《佳木斯大学学报（自然科学版）》2023年第1期175-178,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：2020年度安徽高校自然科学研究重点项目(KJ2020A1008);安徽省2020年度高校学科(专业)拔尖人才学术资助项目(gxbjZD2020118);2020年安徽省高等学校省级质量工程教学研究重大项目(2020jyxm1307)。

摘　　要：由于人体腹腔的特殊构造,肝脏往往在影像学中与其他器官不易辨别。提出一种采用自适应动量估计Adam算法来对肝脏的影像图形进行分析,辅助医疗人员辨别肝脏是否存在病变以及定位存在病灶的确切位置,对于肝脏的影像学分析特别是在影像学图像中的精确识别具有一定的参考与应用价值。Because of the special structure of the human abdominal cavity, the liver is often difficult to distinguish from other organs in imaging. In this paper, we use the Adaptive Moment Estimation(Adam) algorithm combined with U-Net architecture to analyze the medical image of the liver. Moreover, it can assist doctors to identify whether there is a lesion in the liver and locate the exact location of the lesion. Also, it has a certain reference and application value for the imaging analysis of the liver, especially for the accurate recognition of the medical images.

关 键 词：U-Net神经网络 Adam算法 语义分割 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]