A Simple Region Growing Technique for Extraction of ROI from Mammograms
J. Mohanalin, Prem Kumar Kalra, Nirmal Kumar · SSRN Electronic Journal · 2010
In this paper, a region growing algorithm is presented for subsequent use in image segmentation and Region of Interest (ROI) extraction. Breast cancer is one of the leading causes of women’s death in the world. Since the causes are unknown, breast cancer cannot be prevented. Microcalcifications (MCs) are the earliest signs of breast cancer and their detection is one of the most important research areas now. The problem in extracting the MCs alone starts from the presence of radio-opaque artifacts or labels. The presence of these labels makes automation of the MC extraction tougher. This paper presents a simple region growing technique to extract the breast region alone from the mammogram excluding the labels. The advantage of the proposed method is its simplicity when compared to the existing complicated models. Since it does not deal with the breast region, there is no loss of any vital data from the ROI. Many clinical mammograms were used to test the fidelity of this algorithm. The experiments suggest that the breast region is not only accurately extracted but also the computation time is much lesser as the algorithm is less complicated.