Skip to main content
. 2022 Jul 27;9:894503. doi: 10.3389/fcvm.2022.894503

FIGURE 3.

FIGURE 3

Configuration of the UNet architecture to predict the segmentation mask that marks of the regions of the scar, the myocardium and left ventricle (LV) cavity. The input image is a set of 32 frames of cine SAX image frames. The ground truth (output image) consists of an image with four channels, namely, the background (label = 0), the myocardium (label = 1), the LV cavity (label = 2), and the scar (label = 3).