Super-Resolution and Self-Similarity in Magnetic Resonance Imaging

This thesis is about super-resolution reconstruction (SRR) and self-similarity in MRI. These are two overlapping fields of research and in the studies described here, one has naturally lead to the other. From investigating basic properties of conventional approaches to SRR in MRI and applying these methods to specific research problems, we saw a potential improvement to SRR in MRI by employing the selfsimilarity of the images. Self-similarity is a versatile methodology, and beside using it for SRR, we have performed a thorough investigation of its application to voxelwise classification in MRI. In this introductory chapter, we will briefly give some background on SRR and self-similarity in MRI and introduce the five studies included in the thesis.