Researchers are exploring various networks and mechanisms for medical image segmentation, including multiscale attention networks for 3D tumor segmentation in PET images and other applications. Different models, such as dual-branch guided feature interaction networks and transformer-based attention mechanisms, are being developed for tasks like brain tumor segmentation and liver tumor segmentation. These approaches aim to improve image segmentation in medical imaging, with some focusing on specific types of images, like MRI or SPECT/CT.