基于3D ResNet50改进模型的TOF-MRA脑动脉瘤分类方法

薛培阳, 耿辰, 李郁欣, 鲍奕仿, 鲁宇澄, 戴亚康

A Classification Method for Cerebral Aneurysms in TOF-MRA Based on Improved 3D ResNet50 Model

XUE Peiyang, GENG Chen, LI Yuxin, BAO Yifang, LU Yucheng, DAI Yakang

图3 PMAF-Net架构图，上方为网络整体结构，下方A为MDC模块，B、C为CSA模块.残差块底部的数字代表每次下采样后特征图的尺寸和通道数（如323×64表示特征图大小为32×32×32，输出通道数为64）

Fig. 3 PMAF-Net architecture diagram. The top part shows the overall network structure, while the bottom part displays the MDC module (A) and the CSA modules (B and C). The numbers at the bottom of the residual blocks indicate the size and number of channels of the feature map after each downsampling (e.g., 323×64 indicates a feature map size of 32×32×32 with 64 output channels)