Automated Detection of Breast Cancer’s Indicators in Mammogram via Image Processing Techniques

Aims: The detection of abnormalities in mammographic images is an important step in the diagnosis of breast cancer. The indicators of cancer in mammograms can be in form of calcification, mass and stellate lesion. This paper proposed a two-stage procedure for the detection of these cancer’s indicators.
Methodology: Twenty images were used for the study. The images were obtained from Mammographic Image Analysis Society (miniMIAS) database. The images were pre-processed and enhanced using top hat filtering method and the enhanced images were segmented using Otsu’s method. Four features were extracted and selected from the mammographic images using Gray Level Concurrence Matrix (GLCM). The features extracted and selected include energy, homogeneity, contrast, and correlation. Subtractive clustering and fuzzy logic techniques were employed for the classification of the cancer’s indicators in the mammograms. The implementation of the image processing techniques was done with matrix laboratory.
Results: The result showed that seven of the images were affected by stellate lesion, nine of the images were affected by microcalcification while four of the images were affected by mass.
Conclusion: The method presented in this paper would enhance the detection of cancerous cells in the breasts.