Breast cancer is a deadly disease and should be treated at right time to save life from death. Tumor demarcation from digital mammograms is important for classification of benign and malignant masses. Computer aided systems can provide accurate results for breast tumors which can help radiologists in distinguishing malignant masses from benign. Segmentation is based on the principle of grouping similar regions together based on some pre-defined criteria which helps to separate tumor from breast for further processing. Histograms show almost a perfect distribution of intensity levels of pixels in mammograms. This can be specifically used in thresholding process to enhance the tumor region which further simplifies the segmentation methodology. Detection of breast tumors from digital mammograms can be done in many ways. In this paper, Local Thresholding and Mathematical Morphology methodologies are implemented to segment the breast tumor from digital mammograms. Various morphological operators can be used to extract the tumor from breast. In local thresholding image is divided into different parts and thresholding is applied rather than selecting one threshold for the entire image. In this paper both algorithms are implemented and their respective results are presented and compared.

Mathematical Morphology Local Thresholding Digital Mammograms