没有合适的资源?快使用搜索试试~ 我知道了~
人工智能-基于深度改进卷积神经网络的肺结节图像检测.pdf
1.该资源内容由用户上传,如若侵权请联系客服进行举报
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
版权申诉
0 下载量 27 浏览量
2022-06-30
10:45:57
上传
评论
收藏 1.03MB PDF 举报
温馨提示
试读
53页
人工智能-基于深度改进卷积神经网络的肺结节图像检测.pdf
资源推荐
资源详情
资源评论
基于深度改进卷积神经网络的肺结节图像检测
II
实现对原始的肺部 CT 图像进行疑似候选的肺结节的分割与提取。在
搜集到的数据库进行实验,结果表明该方法相比于原模型的检测结果
可有效地将影像中的肺结节检测出来。
(2)通过研究 CT 图像的肺结节检测的分类识别,发现检测识别
的结果中会有较高的假阳性。并通过研究模型融合的方法,发现可以
有效改善假阳性。由此本文采用了模型融合的卷积网络结构的分类模
型。在每次模型训练结束后,引入 boosting 方法,每次将错分类样本
重新分配权重,继续训练模型得到最终的检测结果。实验检测结果表
明了相比于一般的卷积神经网络,提出的模型融合卷积网络可对检测
结果的假阳数进行有效的降低,模型实现了更高的肺结节的检测性
能。
(3)由于医学领域中可用于实际研究的图像的特殊性和数量的稀
缺,所以可获得用于进行研究实验的数据集较少。通过对生成对抗网
络的研究,发现其可以有效的仿造真实样本,从而有效扩充研究对象。
对此本文分析研究了利用生成对抗网络模型用于数据增强的方法来
改善对 CT 图像的肺结节检测模型的性能。该方法首先根据标注信息,
切割肺结节立方块,然后送入 GAN 网络模型训练生成更多的“真结
节”。扩充了一定量的正样本的数据集后,再将其应用到卷积检测模
型中实现对肺结节的检测。以 LIDC 数据库的图像数据作为样本进行
了实验,从检测的结果看,肺结节的检出率可满足实际的要求。
关键词:医学图像处理;卷积神经网络;肺结节检测;生成对抗网络
万方数据
基于深度改进卷积神经网络的肺结节图像检测
III
RESEARCH ON LUNG NODULES IMAGE DETECTION BASED ON
DEEP IMPROVED CONVOLUTIONAL NEURAL NETWORK
ABSTRACT
With the fast development of the society and economy, people's daily lives are increasingly
surrounded by pollution problems. The decline of urban air quality, and frequent foggy weather,
and people's use of tobacco have led to lung cancer becoming a huge threat to human health. The
most effective, and the lowest cost, and low-burden way for the diagnosis and treatment of cancer
is to diagnose and treat lung cancer as soon as possible, and prevents it in advance. The early
manifestations of lung cancer are pulmonary nodules. Compared with other normal organs and
soft tissues, there is no obvious difference. It is mostly present in the lung tissue. In actual medical
diagnosis, it is necessary to use imaging equipment to detect pulmonary nodules. The algorithm of
deep learning is a hot topic of current research and application. Through the deep learning method,
the multi-layer network structure model can be designed and the computer can realize the feature
extraction by itself, which not only reduces the manual participation but also meets the demand for
a large amount of data processing training. The intelligent algorithm of deep learning has been
widely used in many practical fields, and its application in the medical direction has also achieved
good results and benefits for the analysis and processing of medical images. The main content of
this paper is the study of the method of detecting lung nodules in computerized CT images. The
main contents of this paper are as follows:
(1) The lung nodule detection system of CT images was analyzed, and a convolutional
network model with improved network depth was proposed to realize the detection of pulmonary
nodules. Firstly, a simple pre-processing was performed on the lung CT image, and the region of
interest within the image was enhanced. The training data was then used as the input to the model,
and the trained convolutional network model was used to segment and extract the suspected
candidate lung nodules from the original lung CT image. Experiments were carried out in the
万方数据
基于深度改进卷积神经网络的肺结节图像检测
IV
collected database, and the results showed that the method can effectively detect the pulmonary
nodules in the image compared with the original model.
(2) By studying the classification and identification of lung nodule detection in CT images, it
was found that there were higher false positives in the results of detection and recognition. And by
studying the method of model fusion, it was found that it can effectively improve false positives.
Therefore, this paper adopts the classification model of model fusion convolutional network
structure. After training the model one time, a boosting method was introduced to redistribute the
weights of the misclassified samples, and then the model was trained again until the final test
results are obtained. The experimental results show that compared with the general convolutional
neural network, the proposed model fusion convolutional neural network can effectively reduce
the false positive number of the detection results, and the model achieves higher detection
performance of lung nodules.
(3)Due to the particularity and scarcity of medical images that can be used in practical
research, there is little data available for conducting research and experiments. Through the
research on the generation-adveratial network, it is found that it can effectively copy real samples,
thus effectively expanding the number of research objects. Therefore, this paper analyzes the use
of the method of generation-adveratial network model for data enhancement to improve the
performance of lung nodule detection models for CT images. The model can cut the lung nodule
cubes according to the annotation information, which were sent to the network model to generate
more "true nodules". After the number of positive samples is expanded, they are applied to the
convolution detection model to detect lung nodules. Experiments were carried out using the image
data of the LIDC database as training samples. From the results of the test, the detection rate of the
pulmonary nodules can meet the actual requirements.
Xiaoting Wang (Control Science and Engineering)
Supervised by Professor Tao Gong
KEY WORDS: medical image processing; convolutional neural network; lung nodule
detection; generating confrontation network
万方数据
基于深度改进卷积神经网络的肺结节图像检测
V
目录
摘 要.............................................................................................................................. I
ABSTRACT ................................................................................................................. III
第一章 绪论.................................................................................................................. 1
1.1 研究背景与意义.............................................................................................. 1
1.2 国内外研究现状.............................................................................................. 3
1.2.1 基于 CT 图像的肺实质分割算法研究现状 ........................................ 3
1.2.2 基于 CT 图像的肺结节检测算法研究现状 ........................................ 5
1.3 课题研究目的和意义...................................................................................... 7
1.4 本文创新和主要工作...................................................................................... 8
1.5 本文章节安排.................................................................................................. 8
第二章 卷积神经网络的深度改进模型.................................................................... 10
2.1 卷积神经网络深度改进的原因.................................................................... 10
2.2 深度改进卷积神经网络的结构和训练方法................................................ 10
2.2.1 深度改进卷积神经网络的结构.......................................................... 11
2.2.2 基于残差网络的加深.......................................................................... 13
2.2.3 深度改进卷积神经网络的训练方法.................................................. 15
2.3 基于深度改进卷积神经网络的医学图像处理方法.................................... 16
2.4 本章小结........................................................................................................ 17
第三章 基于深度改进卷积神经网络的肺结节候选区域检测算法设计与比较.... 18
3.1 UNET 神经网络的传统肺结节检测算法分析 ............................................. 18
3.2 肺结节 CT 图像的预处理 ............................................................................ 19
3.2.1 肺结节 CT 图像数据的准备 .............................................................. 19
3.2.2 肺部 CT 图像数据归一化处理和肺结节样本数据增强 .................. 20
3.3 基于改进神经网络对肺结节候选区域检测算法的设计与分析................ 20
3.3.1 卷积神经网络深度的改进.................................................................. 20
3.3.2 三种损失函数的误差计算算子设计.................................................. 22
3.4 本算法与传统算法的试验性能对比............................................................ 22
3.4.1 实验性能指标的设计.......................................................................... 22
3.4.2 输入图像块的大小调节...................................................................... 23
3.4.3 两种算法的结果比较分析.................................................................. 24
3.5 本章小结........................................................................................................ 25
万方数据
基于深度改进卷积神经网络的肺结节图像检测
VI
第四章 基于深度改进卷积神经网络的肺结节 CT 图像检测与分类 .................... 26
4.1 传统肺结节 CT 图像检测的卷积神经网络集成学习方法分析 ................ 26
4.2 基于 Boosting 模型和 3DCNN 卷积神经网络的融合改进分析 ................ 27
4.2.1 基于 3DCNN 卷积神经网络的分析 .................................................. 27
4.2.2 基于模型融合的卷积神经网络的分析与改进.................................. 28
4.3 检测候选肺结节图像的预处理.................................................................... 30
4.4 卷积神经网络的优化算法比较分析............................................................ 30
4.4.1 传统卷积神经网络的梯度下降优化算法的分析.............................. 30
4.4.2 Adam 优化算法分析与对比 ................................................................ 32
4.5 实验结果和分析............................................................................................ 32
4.5.1 肺结节检测的分类识别结果展示与分析.......................................... 32
4.5.2 图像预处理对实验结果的影响分析.................................................. 33
4.5.3 模型融合的卷积神经网络和卷积神经网路性能对比...................... 33
4.6 本章小结........................................................................................................ 34
第五章 基于生成对抗网络的肺结节样本的数据增强............................................ 35
5.1 生成对抗网络进行数据增强的原因............................................................ 35
5.2 生成对抗网络的算法分析............................................................................ 36
5.2.1 条件生成对抗网络的模型结构.......................................................... 36
5.2.2 条件生成对抗网络的训练方法.......................................................... 37
5.3 生成对抗网络的图像处理框架.................................................................... 39
5.4 基于生成对抗网络数据增强的卷积神经网络和卷积神经网路性能对比 39
5.5 本章小结........................................................................................................ 40
第六章 总结与展望.................................................................................................... 41
6.1 全文总结........................................................................................................ 41
6.2 展望................................................................................................................ 42
参考文献...................................................................................................................... 44
在校期间主要学术成果.............................................................................................. 50
致谢.............................................................................................................................. 51
万方数据
剩余52页未读,继续阅读
资源评论
programcx
- 粉丝: 40
- 资源: 13万+
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功