Leveraging Anatomical Similarity for Unsupervised Model Learning and Synthetic MR Data Generation

We present a method for the controllable synthesis of 3D (volumetric) MRI data. The model is comprised of three components which are learnt simultaneously from unlabelled data through self-supervision: i) a multi-tissue anatomical model, ii) a probability distribution over deformations of this anatomical model, and, iii) a probability distribution over ‘renderings’ of the anatomical model (where a rendering defines the relationship between anatomy and resulting pixel intensities). After training, synthetic data can be generated by sampling the deformation and rendering distributions.