“Experts propose an end-to-end medical image segmentation framework that combines the advantages of convolutional neural networks and transformers to solve the problems of global local feature fusion and convolutional kernel parameter solidification, thereby improving segmentation accuracy.”