Brain Tumor Segmentation Algorithm Based on Multi-scale Deep Learning Networks

Brain tumors seriously affect people's health, and the brain can be visualized by MRI, which provides the basis for physicians to accurately diagnose. Aiming at the problem that it is difficult to accurately segment brain tumors, a deep learning network is established to realize brain tumor segmentation. Firstly, according to the uncertainty of the region where the brain tumor is located, a multi-core convolutional model is constructed to obtain information at different scales. Then SP_Net is proposed, the constraint function is established and the brain tumor segmentation model is realized.