Skip to main content
. 2020 Feb;37(1):136–141. [Article in Chinese] doi: 10.7507/1001-5515.201903052

图 1.

The structure of 3D U-NET deep convolutional neural network

3D U-NET 深度卷积神经网络结构

conv: a convolution operation is performed; BN: a batch normalization process on the data is performed; Elu: a layer data process using the Elu function is performed; transposed conv: a transposed convolution operation is performed; concat: the two layers are merged

conv:进行一次卷积运算;BN:进行一次批标准化;Elu:使用该激活函数进行一次数据处理;transposed conv:进行一次上采样卷积运算;concat:将两层融合

图 1