Uncertainty evaluation in a fuzzy classifier for microcalcifications in digital mammography

Breast cancer is the most common cancer in women. Fortunately, in the last years, the percentage of recovery has become higher and higher. The reasons are multiple: prevention, better surgery techniques, better diagnoses. In order to improve diagnostic performance, many Computer Aided Diagnosis systems have been developed in the latest years, which help the radiologists to detect and classify lump-masses. These systems, however, do not yet consider the uncertainty associated to the measurements (the mammography), even if, from a metrological point of view, they should. In this paper, measurement uncertainty is considered through a fuzzy inference system for the classification of microcalcifications in digital mammography.