C-3D可变形卷积神经网络模型的肺结节检测 下载: 1104次
Detection of Pulmonary Nodules Based on C-3D Deformable Convolutional Neural Network Model
1 上海海洋大学工程学院, 上海 210306
2 上海建桥学院机电学院, 上海 210306
3 上海市第六人民医院东院放射介入科, 上海 210306
图 & 表
图 1. 肺部CT扫描图
Fig. 1. CT images of lung
下载图片 查看原文
图 2. C-3D卷积神经网络模型结构
Fig. 2. Structure of C-3D convolutional neural network model
下载图片 查看原文
图 3. 三维可变形卷积与池化结构。(a)三维可变形卷积结构;(b)三维可变形池化结构
Fig. 3. Structures of 3D deformable convolution and pooling. (a) 3D deformable convolution; (b) 3D deformable pooling
下载图片 查看原文
图 4. C-3D可变形卷积神经网络模型结构
Fig. 4. Structure of C-3D deformable convolutional neural network model
下载图片 查看原文
图 5. 不同学习率和优化函数下的分类正确率。(a)学习率实验对比;(b)优化函数实验对比,
Fig. 5. Classification accuracy for different learning rates and optimization functions. (a) Experimental comparison of learning rate; (b) experimental comparison of optimization function
下载图片 查看原文
图 6. 不同模型的ROC曲线图和PRC曲线图。(a) ROC曲线;(b) PRC曲线
Fig. 6. ROC curves and PRC curves of different models. (a) ROC curves; (b) PRC curves
下载图片 查看原文
图 7. 可变形卷积层中不同特征图数量的箱型图。(a) AUC指标箱型图;(b) F1指标箱型图;(c) P指标箱型图;(d) Recall指标箱型图
Fig. 7. Boxes of deformable convolution layers with different numbers of features. (a) Box of AUC; (b) box of F1; (c) box of P; (d) box of R
下载图片 查看原文
图 8. 卷积窗口像素采样可视化结果。(a)原始的肺结节标注图片;(b)原C-3D卷积神经网络模型;(c)改进后的C-3D卷积神经网络模型
Fig. 8. Visualization results of convolution window pixel sampling. (a) Original labeled lung images; (b) original C-3D convolutional neural network; (c) improved C-3D convolutional neural network
下载图片 查看原文
表 1C-3D可变形卷积神经网络模型中卷积层与池化层的具体参数设置
Table1. Setting of parameters of convolution layer and pooling layer in C-3D deformable convolutional neural network
Layer | F (number of convolutionalfeature maps)×S (kernal size) | Stride | Activation function | Output |
---|
3D conv1 | 16×(3×3×3) | 1×1×1 | ReLU | 32×32×32 | 3D conv2 | 64×(3×3×3) | 2×1×1 | ReLU | 16×32×32 | 3D pool1 | 64×(1×2×2) | 1×2×2 | — | 16×16×16 | 3D conv3 | 128×(3×3×3) | 1×1×1 | ReLU | 16×16×16 | 3D pool2 | 128×(2×2×2) | 2×2×2 | — | 8×8×8 | 3D conv4 | 256×(3×3×3) | 1×1×1 | ReLU | 8×8×8 | 3D pool3 | 256×(2×2×2) | 2×2×2 | — | 4×4×4 | 3D conv5 | 512×(3×3×3) | 1×1×1 | ReLU | 4×4×4 | 3D pool4 | 512×(2×2×2) | 2×2×2 | — | 2×2×2 | 3D conv6 | 64×(3×3×3) | 1×1×1 | ReLU | 1×1×1 | 3D conv7 | 1×(1×1×1) | 1×1×1 | Sigmoid | 1×1×1 | Fc (fully connected layer) | — | — | — | 1 |
|
查看原文
表 2数据集信息
Table2. Dataset information
Number | Name | Number ofsamples | Number ofpulmonarynodules |
---|
1 | LIDC-IDIR | 888 | 1186 | 2 | East Hospital of ShanghaiSixth People's Hospital | 300 | 346 | 3 | Hospital of KanghuaHaining Zhejiang | 200 | 212 |
|
查看原文
表 3不同卷积神经网络模型的结果对比
Table3. Comparison of results of different convolutional neural network models
Model | AUC | R | P | F1 |
---|
C-3D | 0.9422±0.0105 | 0.9125±0.0131 | 0.9335±0.0135 | 0.8846±0.0036 | C-3D±DC | 0.9513±0.0127 | 0.9228±0.0158 | 0.9387±0.0101 | 0.8978±0.0097 | C-3D±DP | 0.9326±0.0241 | 0.8997±0.0179 | 0.9254±0.0157 | 0.8797±0.0103 | C-3D±DCP | 0.9575±0.0098 | 0.9183±0.0096 | 0.9413±0.0075 | 0.9067±0.0058 |
|
查看原文
表 4C-3D可变形卷积神经网络中不同输出特征图结果对比
Table4. Results comparison of different output features in C-3D deformable convolutional neural network
Model | R | P | F1 | AUC |
---|
C-3D+DCP,F:8 | 0.9131±0.0125 | 0.9226±0.0048 | 0.8948±0.0107 | 0.9394±0.0117 | C-3D±DCP,F:16 | 0.9183±0.0096 | 0.9413±0.0075 | 0.9067±0.0058 | 0.9575±0.0098 | C-3D±DCP,F:32 | 0.9253±0.0164 | 0.9534±0.0104 | 0.9286±0.0083 | 0.9617±0.0131 | C-3D±DCP,F:64 | 0.8957±0.0063 | 0.9113±0.0243 | 0.9024±0.0096 | 0.9339±0.0164 |
|
查看原文
表 5各个卷积神经网络算法比较
Table5. Comparison of different convolutional neural networks
Model | AUC | P | R | F1 | Data |
---|
Ref. [6] | 0.8434 | 0.7444 | 0.7916 | 0.7673 | LIDC-IDRI | Ref. [7] | 0.9387 | 0.8750 | 0.9243 | 0.8990 | LIDC-IDRI | Ref. [8] | 0.8523 | 0.9045 | 0.9162 | 0.9103 | LIDC-IDRI | Proposed method | 0.9617 | 0.9534 | 0.9239 | 0.9286 | LIDC-IDRI | Proposed method | 0.9661 | 0.9613 | 0.9253 | 0.9378 | LIDC-IDIR+Local data |
|
查看原文
阮宏洋, 陈志澜, 程英升, 杨凯. C-3D可变形卷积神经网络模型的肺结节检测[J]. 激光与光电子学进展, 2020, 57(4): 041013. Hongyang Ruan, Zhilan Chen, Yingsheng Cheng, Kai Yang. Detection of Pulmonary Nodules Based on C-3D Deformable Convolutional Neural Network Model[J]. Laser & Optoelectronics Progress, 2020, 57(4): 041013.