TL-MSE-Net: Transfer learning based nested model for cerebrovascular segmentation with aneurysms

Chaoran Zhang, Ming Zhao, Yixuan Xie, Rui Ding, Ming Ma, Kaiwen Guo, Hongzhen Jiang, Wei Xi, Likun Xia

Abstract: Highlights•A transferability framework is designed to address the small amount of private clinical cerebrovascular datasets.•Exploration of adaptive strategies for different modalities of cerebrovascular datasets (MRI-CT).•Extracting multiscale features for the identification of edge voxels in thin cerebral vessels.•We have validated the effectiveness of our model in cerebrovascular segmentation by using three datasets.