Recently, machine learning methods and particularly deep learning achieved excellent performance in many tasks. Many technologies from autonomous driving to face detection available in smartphone cameras are using these algorithms. Convolutional neural networks (CNNs) is the most popular network in image processing, classification, and segmentation tasks. This network shows high performance in medical image segmentation. However, the results are optimal in scans with artifacts (see the second column of the above figure). These scans visualize motion artifacts that are occurring due to infants' movement during the scanning. To overcome this challenge, an algorithm using another type of neural network recently widely used in image generation called generative adversarial network (GAN) is used. GAN composed of two networks, a generator, and a discriminator. We used this algorithm to remove motion artifacts from scans and improve brain tissue segmentation simultaneously. The results are published in MICCAI 2019 and can be find here.