文献类型：会议论文

英文题名：A novel method for lung masses detection and location based on deep learning

作者：Li, Zirong[1,2];Li, Lian[1]

第一作者：Li, Zirong;李子荣

通信作者：Li, ZR[1];Li, ZR[2]

机构：[1]Lanzhou Univ, Sch Informat Sci & Engn, Lanzhou, Gansu, Peoples R China;[2]Gansu Univ Chinese Med, Sch Informat & Engn, Lanzhou, Gansu, Peoples R China

第一机构：Lanzhou Univ, Sch Informat Sci & Engn, Lanzhou, Gansu, Peoples R China

通信机构：[1]corresponding author), Lanzhou Univ, Sch Informat Sci & Engn, Lanzhou, Gansu, Peoples R China;[2]corresponding author), Gansu Univ Chinese Med, Sch Informat & Engn, Lanzhou, Gansu, Peoples R China.|[10735]甘肃中医药大学;

会议论文集：Biological Ontologies and Knowledge Bases Workshop at IEEE International Conference on Bioinformatics and Biomedicine (IEEE BIBM)

会议日期：NOV 13-16, 2017

会议地点：Kansas City, MI

语种：英文

外文关键词：deep learning; lung masses; Faster RCNN; RESNET; chest radiography database

年份：2017

摘要：As the diagnosis of lung cancer, lung mass for the diagnosis of the disease is meaningful, chest radiography has low price, low radiation, popularity and other characteristics, it is a significant attempt for the location of chest masses on chest radiography using deep learning method. In this paper we have established a labeled lung mass database, and presented a state of the art deep learning methodology for classifying, detecting and locating lung masses on the database. Moreover we analyzed the details of the Faster RCNN network and its architecture, and studied the feature extraction parts by two different networks, both of them are deep learning method. To a certain extent, the two networks can locate the masses. We find that the methodology using RESNET for feature extraction is more satisfying than VGG16, the Ap achieved 52.38% by comparing the test results. The system retrieved 41 out of 51 masses in the testing phase.