MR Slice Profile Estimation by Learning to Match Internal Patch Distributions

To super-resolve the through-plane direction of a multi-slice 2D magnetic resonance (MR) image, its slice selection profile can be used as degeneration model from high resolution (HR) to low (LR) create paired data when training supervised algorithm. Existing super-resolution algorithms make assumptions about since it is not readily known for given image. In this work, we estimate specific image by learning match internal patch distributions. Specifically, assume that after applying correct profile, distribution along HR in-plane directions should LR direction. Therefore, incorporate estimation part generator in generative adversarial network (GAN). way, learned without any external data. Our algorithm was tested using simulations isotropic MR images, incorporated demonstrate benefits, and also tool measure resolution. code at https://github.com/shuohan/espreso2 .