基于YOLOv3深度学习的有限数据集肺结节检测

Q4 Biochemistry, Genetics and Molecular Biology Molecular & Cellular Biomechanics Pub Date : 2022-01-01 DOI:10.32604/mcb.2022.018318
Zhaohui Bu, Xuejun Zhang, Jianxiang Lu, Huan Lao, Chan Liang, Xianfu Xu, Yini Wei, Hongjie Zeng
{"title":"基于YOLOv3深度学习的有限数据集肺结节检测","authors":"Zhaohui Bu, Xuejun Zhang, Jianxiang Lu, Huan Lao, Chan Liang, Xianfu Xu, Yini Wei, Hongjie Zeng","doi":"10.32604/mcb.2022.018318","DOIUrl":null,"url":null,"abstract":"The early symptom of lung tumor is always appeared as nodule on CT scans, among which 30% to 40% are malignant according to statistics studies. Therefore, early detection and classification of lung nodules are crucial to the treatment of lung cancer. With the increasing prevalence of lung cancer, large amount of CT images waiting for diagnosis are huge burdens to doctors who may missed or false detect abnormalities due to fatigue. Methods: In this study, we propose a novel lung nodule detection method based on YOLOv3 deep learning algorithm with only one preprocessing step is needed. In order to overcome the problem of less training data when starting a new study of Computer Aided Diagnosis (CAD), we firstly pick up a small number of diseased regions to simulate a limited datasets training procedure: 5 nodule patterns are selected and deformed into 110 nodules by random geometric transformation before fusing into 10 normal lung CT images using Poisson image editing. According to the experimental results, the Poisson fusion method achieves a detection rate of about 65.24% for testing 100 new images. Secondly, 419 slices from common database RIDER are used to train and test our YOLOv3 network. The time of lung nodule detection by YOLOv3 is shortened by 2–3 times compared with the mainstream algorithm, with the detection accuracy rate of 95.17%. Finally, the configuration of YOLOv3 is optimized by the learning data sets. The results show that YOLOv3 has the advantages of high speed and high accuracy in lung nodule detection, and it can access a large amount of CT image data within a short time to meet the huge demand of clinical practice. In addition, the use of Poisson image editing algorithms to generate data sets can reduce the need for raw training data and improve the training efficiency.","PeriodicalId":48719,"journal":{"name":"Molecular & Cellular Biomechanics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Lung Nodule Detection Based on YOLOv3 Deep Learning with Limited Datasets\",\"authors\":\"Zhaohui Bu, Xuejun Zhang, Jianxiang Lu, Huan Lao, Chan Liang, Xianfu Xu, Yini Wei, Hongjie Zeng\",\"doi\":\"10.32604/mcb.2022.018318\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The early symptom of lung tumor is always appeared as nodule on CT scans, among which 30% to 40% are malignant according to statistics studies. Therefore, early detection and classification of lung nodules are crucial to the treatment of lung cancer. With the increasing prevalence of lung cancer, large amount of CT images waiting for diagnosis are huge burdens to doctors who may missed or false detect abnormalities due to fatigue. Methods: In this study, we propose a novel lung nodule detection method based on YOLOv3 deep learning algorithm with only one preprocessing step is needed. In order to overcome the problem of less training data when starting a new study of Computer Aided Diagnosis (CAD), we firstly pick up a small number of diseased regions to simulate a limited datasets training procedure: 5 nodule patterns are selected and deformed into 110 nodules by random geometric transformation before fusing into 10 normal lung CT images using Poisson image editing. According to the experimental results, the Poisson fusion method achieves a detection rate of about 65.24% for testing 100 new images. Secondly, 419 slices from common database RIDER are used to train and test our YOLOv3 network. The time of lung nodule detection by YOLOv3 is shortened by 2–3 times compared with the mainstream algorithm, with the detection accuracy rate of 95.17%. Finally, the configuration of YOLOv3 is optimized by the learning data sets. The results show that YOLOv3 has the advantages of high speed and high accuracy in lung nodule detection, and it can access a large amount of CT image data within a short time to meet the huge demand of clinical practice. In addition, the use of Poisson image editing algorithms to generate data sets can reduce the need for raw training data and improve the training efficiency.\",\"PeriodicalId\":48719,\"journal\":{\"name\":\"Molecular & Cellular Biomechanics\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Molecular & Cellular Biomechanics\",\"FirstCategoryId\":\"1087\",\"ListUrlMain\":\"https://doi.org/10.32604/mcb.2022.018318\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"Biochemistry, Genetics and Molecular Biology\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Molecular & Cellular Biomechanics","FirstCategoryId":"1087","ListUrlMain":"https://doi.org/10.32604/mcb.2022.018318","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"Biochemistry, Genetics and Molecular Biology","Score":null,"Total":0}
引用次数: 2

摘要

肺肿瘤的早期症状在CT扫描上多表现为结节,据统计,其中30% ~ 40%为恶性。因此,早期发现和分类肺结节对肺癌的治疗至关重要。随着肺癌患病率的增加,大量等待诊断的CT图像给医生带来了巨大的负担,医生可能会因疲劳而漏诊或误检异常。方法:在本研究中,我们提出了一种基于YOLOv3深度学习算法的肺结节检测新方法,只需要一个预处理步骤。为了克服计算机辅助诊断(CAD)新研究开始时训练数据较少的问题,我们首先选取少量病变区域,模拟有限数据集的训练过程:选择5个结节模式,通过随机几何变换变形为110个结节,然后使用泊松图像编辑融合到10个正常肺CT图像中。实验结果表明,泊松融合方法对100幅新图像的检测率约为65.24%。其次,使用来自公共数据库RIDER的419个切片来训练和测试YOLOv3网络。与主流算法相比,YOLOv3对肺结节的检测时间缩短了2-3倍,检测准确率为95.17%。最后,利用学习数据集对YOLOv3的配置进行优化。结果表明,YOLOv3在肺结节检测方面具有速度快、准确率高的优点,可在短时间内获取大量CT图像数据,满足临床实践的巨大需求。此外,使用泊松图像编辑算法生成数据集可以减少对原始训练数据的需求,提高训练效率。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Lung Nodule Detection Based on YOLOv3 Deep Learning with Limited Datasets
The early symptom of lung tumor is always appeared as nodule on CT scans, among which 30% to 40% are malignant according to statistics studies. Therefore, early detection and classification of lung nodules are crucial to the treatment of lung cancer. With the increasing prevalence of lung cancer, large amount of CT images waiting for diagnosis are huge burdens to doctors who may missed or false detect abnormalities due to fatigue. Methods: In this study, we propose a novel lung nodule detection method based on YOLOv3 deep learning algorithm with only one preprocessing step is needed. In order to overcome the problem of less training data when starting a new study of Computer Aided Diagnosis (CAD), we firstly pick up a small number of diseased regions to simulate a limited datasets training procedure: 5 nodule patterns are selected and deformed into 110 nodules by random geometric transformation before fusing into 10 normal lung CT images using Poisson image editing. According to the experimental results, the Poisson fusion method achieves a detection rate of about 65.24% for testing 100 new images. Secondly, 419 slices from common database RIDER are used to train and test our YOLOv3 network. The time of lung nodule detection by YOLOv3 is shortened by 2–3 times compared with the mainstream algorithm, with the detection accuracy rate of 95.17%. Finally, the configuration of YOLOv3 is optimized by the learning data sets. The results show that YOLOv3 has the advantages of high speed and high accuracy in lung nodule detection, and it can access a large amount of CT image data within a short time to meet the huge demand of clinical practice. In addition, the use of Poisson image editing algorithms to generate data sets can reduce the need for raw training data and improve the training efficiency.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Molecular & Cellular Biomechanics
Molecular & Cellular Biomechanics CELL BIOLOGYENGINEERING, BIOMEDICAL&-ENGINEERING, BIOMEDICAL
CiteScore
1.70
自引率
0.00%
发文量
21
期刊介绍: The field of biomechanics concerns with motion, deformation, and forces in biological systems. With the explosive progress in molecular biology, genomic engineering, bioimaging, and nanotechnology, there will be an ever-increasing generation of knowledge and information concerning the mechanobiology of genes, proteins, cells, tissues, and organs. Such information will bring new diagnostic tools, new therapeutic approaches, and new knowledge on ourselves and our interactions with our environment. It becomes apparent that biomechanics focusing on molecules, cells as well as tissues and organs is an important aspect of modern biomedical sciences. The aims of this journal are to facilitate the studies of the mechanics of biomolecules (including proteins, genes, cytoskeletons, etc.), cells (and their interactions with extracellular matrix), tissues and organs, the development of relevant advanced mathematical methods, and the discovery of biological secrets. As science concerns only with relative truth, we seek ideas that are state-of-the-art, which may be controversial, but stimulate and promote new ideas, new techniques, and new applications.
期刊最新文献
Hot Topics of Molecular and Cellular Biomechanics in 2022 CFD Study on Hemodynamic Characteristics of Inferior Vena Cava Filter Affected by Blood Vessel Diameter Can PAPE-Induced Increases in Jump Height Be Explained by Jumping Kinematics? Reconstruction of the Hindlimb Locomotion of Confuciusornis (Aves) and Its Implication for the Origin of Avian Flight Classification of Leukemia and Leukemoid Using VGG-16 Convolutional Neural Network Architecture
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1