Characterization of the Normal Mammograms Based on Statistical Features

Abstract

Abstract— In the mammographic images, the textural features of the normal tissue are identical what so ever the type of the abnormalities that may exist or not on the mammogram. In this article we present a new method for the characterization of normal mammograms. Assuming randomness of the breast tissues, we calculate from the mammographic image the cooccurrence matrix and Harralick parameters. Tested on a data base composed of 30 normal mammograms, the method revealed that among the set of parameters calculated, the variance, the contrast and the correlation represent the most significant features in characterizing both the fat tissue and the mammary gland that constitute the tissue of a normal breast.

Cite this paper

@inproceedings{Eddaoudi2006CharacterizationOT, title={Characterization of the Normal Mammograms Based on Statistical Features}, author={F . Eddaoudi and F. Regragui and K . Laraki}, year={2006} }