Pixel-Label-Based Segmentation of Cross-Sectional Brain MRI Using Simplified SegNet Architecture-Based CNN

Figure 3

Representation of the proposed method, the training and testing image is separated, so that training images after preprocessing feeds into CNN network along with its respective ground truth, after reaching to convergence, the training is stopped and the network is now called trained. This trained network is used to test other test image separately to get segmented image, which is compared with its ground truth itself for performance analysis.