TY - JOUR T1 - Automatic Segmentation of Breast Mammograms Using Hybrid Density Slicing and k-mean Adaptive Methods AU - Ibrahim M. Ali, Semaa AU - H. Salman, Nassir JO - Journal of Engineering and Applied Sciences VL - 14 IS - 15 SP - 5044 EP - 5050 PY - 2019 DA - 2001/08/19 SN - 1816-949x DO - jeasci.2019.5044.5050 UR - https://makhillpublications.co/view-article.php?doi=jeasci.2019.5044.5050 KW - Mammograms KW -breast cancer KW -image segmentation KW -image processing KW -adaptive KW -k-means and density slicing AB -

Medical imaging is a fundamental theme of contemporary healthcare and its engineers take mammograms, ultrasounds, X-rays and computed tomography images to analyze patient’s hurts and illnesses. Segmenting the mammogram into diverse mammographic densities is strategic for risk evaluation and measurable appraisal of density variations to extract the cancer regions. Accordingly in this study, the application of density slicing and k-mean adaptive techniques has been conducted to explore the boundary of changed breast tissue areas in mammograms. The objective of the segmentation process is to perceive if density slicing and k-mean adaptive procedure have the feasibility split diverse densities for the diverse breast outlines. The density slicing is used to make available hard limitation while the thresholds are designated in accordance with user-defined and radiology. k-mean adaptive has been used to cluster the region where the initial seed was based on the mean of array multiply by 0.05. Density slicing has processed on images of numerous imaging modalities without mammograms consideration. As a result, this study is for all intents and purposes concentrated on using hybrid method of density slicing and k-mean adaptive process to achieve segmentation to augment the discernibility of diverse breast densities in mammography images. The suggested approach for the segmentation of mammograms on the source of their region into diverse densities based classifications has been investigated on mini-MIAS database. The concluding consequences show instinctive segmentation of ROI with edge map and dissimilar properties extraction for the investigation process.

ER -