Medical Image Registration: interpolations, similarities and optimizations strategies

This paper presents a study conducted for evaluating differentinterpolation schemes, similarity metrics and optimizationalgorithms for the purpose of volumetric medicalimage registration. Each technique has been implementedto be plugged in a modular system. Rotation, translationand scale error has been measured to obtain a performanceevaluation for all of the combinations of the consideredtechniques. Several experimental tests were conducted forvalidation both on synthetic and real datasets providing anexhaustive overview of the various strategies used.