Automatic Detection of Microcalcifications in Mammograms using Genetic Algorithm

Thangavel Kuttiannan (1), Karnan Marcus (2)
1: Department of Mathematics, Gandhigram Rural Institute-Deemed University,
2: Department of computer science, Gandhigram Rural Institute-Deemed University, Gandhigram-624302, Tamil Nadu, India.

Microcalcification on x-ray mammogram is a significant mark for early detection of breast cancer. The bilateral subtraction techniques applied to mammogram has very important practical applications in computerized detecting asymmetries (microcalcification, architectural distortions, etc.) between the left and right breast images. In this paper, a computer aided diagnosis system for the automatic finding of microcalcification from digitized mammograms is developed using bilateral subtraction. The pectoral muscle region is removed from the mammogram image as a prerequisite for detection of breast border. The proposed system consists of two steps: First step is detection of the breast border and nipple on digital mammograms. To detect the breast border, an algorithm that computes the gradient of gray levels, average gradient method, logarithmic transformation, and genetic algorithm are applied. Second, using the border and the nipple position of left and right mammograms; the images are correlated and subtracted to extract the suspicious regions. Before subtraction the mammogram images are normalized. Four different algorithms are identifying the nipple position (maximum height of the breast border, maximum gradient, and maximum second derivative of the gray levels across the median-top section of the breast and genetic algorithm hybrid with second derivative). Computerized determination of breast border and nipple allows the alignment of right and left breast images to subtract. The algorithms are tested on 161 pairs of digitized mammograms from MIAS database and 125 pairs of mammograms obtained from various hospitals.

Key words: Breast boarder, nipple identification, genetic algorithm, bilateral subtraction, asymmetries.