Image understanding methods applied to medical diagnostics

This chapter presents the practical operation of methods, algorithms and techniques described in previous ones. They will be presented using two selected examples. In previous papers and books (for example [34, 36, 37, 38, 39, 55]), we showed the applicability and proved the usefulness of the automatic understanding technology in some medical applications (e.g. pancreatic cancer or ureter stenoses), but all such examples were related to the organs presented in medical images as singular, convex, two-dimensional shapes. In this book, we try to extend the evaluation of the considered technologies to cover more complicated medical images. Specifically, we will present and discuss a new technique of computer-aided analysis and recognition of pathological wrist bone lesions. In this application, the automatic understanding technology must prove its utility for a complex anatomical structure (the hand), formed of many separate elements (wrist bones), where the pathologies considered can be represented not only by deformations of the forms, dimensions and shapes of particular selected bones, but can also be expressed in pathological relationships between individual bones. This problem is much more interesting from the scientific point of view, because automatic understanding is much more complicated when it is based on relationships between elements instead of analyzing one particular morphological element itself.