Automatic Detection of Breast Lesion Contour and Analysis using Fractals through Spectral Methods

Lesions and its contours are prominent signatures to determine malignancy in mammograms. Detection of the masses and their spread in mammogram is important for radiologists. It is also important to detect the shape of the contour or boundary to delineate malignant and benign lesions as malignant lesions have speculated or ill-defined boundary and benign mass have smooth boundary. Automatic detection of boundary helps the doctors in analyzing the lesion in less time and prevents unnecessary biopsies. In this paper we proposed algorithms for 1) Image enhancement using homomorphic filtering and adaptive histogram equalization technique 2) Segmentation using Enhanced K means clustering 3) Contour Extraction using morphological operations 4) Fractal analysis of the signatures of contours using Power spectra 5) Extraction geometric features from the lesions. These algorithms have been tested on 34 mammograms.