基于全卷积网络的乳腺肿瘤动态增强磁共振图像分割

邱玥,聂生东,魏珑

Segmentation of Breast Tumors Based on Fully Convolutional Network and Dynamic Contrast Enhanced Magnetic Resonance Image

Yue QIU,Sheng-dong NIE,Long WEI

表1 CBP5-Net结构

Table 1 The structure of CBP5-Net

结构 输入 输出 结构 输入 输出 block1 conv 1282*1 1282*32 block4 conv 162*128 162*64 BN 1282*32 1282*32 BN 162*64 162*64 max-pooling 1282*32 642*32 max-pooling 162*64 82*64 block2 conv 642*32 642*64 block5 conv 82*64 82*32 BN 642*64 642*64 BN 82*32 82*32 max-pooling 642*64 322*64 max-pooling 82*32 42*32 block3 conv 322*64 322*128 flatten 42*32 512 BN 322*128 322*128 dense 512 64 max-pooling 322*128 162*128 dense 64 2