HIGHLIGHTS
- who: Adamu and collaborators from the School of Microelectronicss, Tianjin University, Tianjin, China have published the research: HMNet: Hierarchical Multi-Scale Brain Tumor Segmentation Network, in the Journal: (JOURNAL)
- what: The authors propose a novel segmentation network (HMNet) which contains a high-resolution branch and parallel multi-resolution branches. To overcome the large computational overhead caused by expensive 3D convolution the authors propose a lightweight conditional channel weighting block to reduce GPU memory and improve the efficiency of Attention mechanisms are often introduced in brain tumor segmentation networks to make the model more focused on . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.