U-Net-based Semi-supervised learning Transformer for the segmentation of Epicardial Adipose Tissue (EAT)

A small amount of fat called epicardial adipose tissue (EAT) surrounds the heart muscle. Since differences in volume and size are closely related to cardiovascular problems, EAT segmentation is extremely important among those who work in the cardiovascular field. The complexity of this segmentation process, however, makes it more difficult. Several deep-learning techniques have been utilized successfully to allay this worry. However, to segment an image effectively, one must have a thorough understanding of its global information as well as its position and spatial information. The unique U-Net semi-supervised transformer network presented in this paper specifically aims to accomplish this. The segmentation procedure was much improved with the method. The segmentation procedure was much improved with the method. Using Coronary Computed Tomography Angiography(CCTA) data, the suggested model is assessed with average dice scores of above 92%.