U_Net_Ni_WC_Optical_images

This data was used to train a U-Net convolutional neural network to semantically segment WC from Ni-WC optical microscopy images. The training and test datasets contain 194 and 37 512x512 pixel images and their corresponding ground truths, respectively.