Single level UNet3D with multipath residual attention block for brain tumor segmentation

Atrous convolution and attention have improved the performance of the UNet architecture for segmentation purposes. However, a perfect combination of atrous convolution and attention to improve brain tumor segmentation performance is still an interesting challenge. In this paper, we propose UNet arch... View More