deepbrain
deepbrain copied to clipboard
How does the segmentation work?
Hi I just want to know how does segmentation work?? I cannot find the explanation for the segmentation code here
It uses a convolutional neural network based on U-Net. The code to this network is here. This network was trained with already segmented brain images ( CC359 dataset, NFBS dataset and ADNI dataset). The result of this training is this file with the network weights and biases. To segment, it's necessary to load this file with its weights and biases, pass the image you want to segment. The network will return to you a probability of each image voxel being part of the brain. Just select the voxels with propability larger than some p (by default 0.5).