A mathematical representation of biological variability in medical images

Larisa Matejic
2003 Quarterly of Applied Mathematics  
Medical image ensembles exhibit variability and it is the aim of computational anatomy to represent such variabilities mathematically and to exploit these knowledge representations by inference algorithms implemented through code. The variability is caused by several factors and our pattern theoretic approach rests on the assumption that they can be understood in terms of groups of transformations and probability measures on such groups. We shall arrange the similarity groups in a cascade,
more » ... in a cascade, typically starting with the more rigid transformations and continuing with more flexible ones. Most importantly, however, we attach great significance to the physical and biological interpretation of the similarity groups.
doi:10.1090/qam/1955221 fatcat:y7hxiibpcnbirb5b5hbh3daz54