Detection of Breast Cancer Tumor based on Morphological Watershed Algorithm

 

H.S.Sheshadri and A. Kandaswamy
Department of ECE, PSG College of Technology, Coimbatore-641004

 

Abstract:

Mammograms are the soft X-rays meant for the detection of any lesions or cysts in breasts. Now days it is playing major role in the detection of breast cancer at an early stage. In this paper we have developed a lesion segmentation technique based on the extraction of catchments basins through a topographic representation of the mammography image. Applying the watershed algorithm on a rough image leads to over-segmentation problems. To avoid this, we have first carried out a preprocessing step which consists to remove or attenuate the curvilinear structures present in a mammogram and corresponding to the blood vessels, veins, milk ducts, speculations and fibrous tissue. Then the gradient of the preprocessed image is calculated and finally the segmentation algorithm is applied to this latter image. We have tested this proposed method on digital mammograms taken from the mini-MIAS data base and found that the lesion segmentation algorithm proposed closely match radiologists outlines of these lesions.

 

Keywords: Mammogram, Texture analysis, Cancer detection, morphological theory, segmentation, watershed algorithm, filtering

 

BibTex: