An automatic and simple breast tumor classification using area matching

This paper proposes a simple and automatic method for classification of breast malignancy in digital mammogram images. The method is based on circular approximation of the contour extracted from the mammogram image. Then it matches the overlapping area between the tumor's contour and its circular approximation. The fitting circle is obtained by considering the centroid of tumor as its centre and the arithmetic mean of maximum and minimum radial distances of contour points from the centroid, as its radius. The similarity between the fitted circle and tumor is measured in terms of area matching as a feature for classification of tumor as malignant or benign. The simulation results show that for a set of 150 tumor contours, the performance obtained in terms of the receiver operating characteristic (ROC) parameters like accuracy (A
_c
), sensitivity (S
_e
), specificity (S
_p
), and positive (PPV) and negative predictive values (NPV) are 94%, 0.9494, 0.9296, 0.9375 and 0.9429 respectively.