FAST MAMMOGRAM SEGMENTATION ALGORITHM FOR SEGMENTING FIBROGLANDULAR TISSUE

Screening mammography is considered one of the most effective ways for the early detection of breast cancer. Previous works on breast tissue identification and abnormalities detection notice that the feature extraction process is affected if the region processed is not well focused. Thereby, it is important to split the mammogram into interesting regions to achieve optimal breast parenchyma measurements, breast registration or to put into focus a technique when we search for abnormalities. In this paper, an automated technique for segmenting a digital mammogram to extract glandular tissue. This algorithm is divided into four main phases. In the first phase, shrinking and reducing image mammogram. Secondly, extract breast region from background and pectoral muscle has suppressed. Thirdly, image histogram analysis based on the concept of minimum cross-entropy is used to calculate an optimal threshold that gives a preliminary glandular segmentation, fourthly, it is designed to improve the quality of segmentation. This improvement is comprised of a set ofpost-processing operations (segment elongated region and applying morphological operation).