An algorithm for image segmentation and its applications in medical imaging.

Segmentation of medical images is required to obtain geometrical measures. There are two main classes of segmentation algorithms. The first one requires a seed or a selection performed by a trained operator while the second one does not require operator intervention. The algorithm object of this work belongs to the second class. The distribution of the signal in connected points in a region of a given dimension is used to classify any point as belonging to a ROI or another ROI minimizing a cost function. The result of the elaboration is a new image partitioned in ROIs.