The presence of clusters of microcalciﬁcations in mam- mograms is particularly signiﬁcant for early detection of breast cancer. In this paper a Computer Aided Detection system designed for this task is described. The detection of microcalciﬁcations is performed by means of a segmen- tation based on a watershed transform and a further anal- ysis based both on heuristic rules and AdaBoost classiﬁ- cation. Finally a clustering algorithm is applied to detect those clusters of medical interest. The approach has been successfully tested on a Full Field Digital Mammographic database that has been developed through a strong cooper- ation between radiologists and computer scientists.
File in questo prodotto:
Non ci sono file associati a questo prodotto.