A Fuzzy Approach of the Correspondence of Microcalcifications

Maria Rifqi (1) , Jean-Yves Delort (1) , Bernadette Bouchon-Meunier (1)

1 APA - Apprentissage et Acquisition des connaissances

This paper addresses the correspondence problem in fuzzy image processing. This problem is a generic one in computer vision, arising whenever a match must be found between the features of two related patterns. The final goal of our problem is to propose an aid when a biopsy using stereotactic guidance is used to confirm the diagnostic of a breast cancer. This implies a matching process of microcalcifications of two views. The matching process of microcalcifications is divided into four steps: the detection of all microcalcifications on the two views (left and right); the construction of a correspondence matrix which indicates all the possible combinations of microcalcifications; the construction of a resemblance matrix, and the exploitation of this matrix. This paper focuses on the two last steps of the process. We propose an approach that uses the general framework of comparison of fuzzy values and the general assignment problem.

Domaines: Informatique [cs]; Intelligence artificielle [cs.AI]

Maria Rifqi, Jean-Yves Delort, Bernadette Bouchon-Meunier. A Fuzzy Approach of the Correspondence of Microcalcifications. EUropean Network on Intelligent TEchnologies for Smart Adaptive Systems Conference, EUNITE 2001, Dec 2001, Tenerife, Spain. ⟨hal-01075375⟩