Novel approach to locate region of interest in mammograms for Breast cancer.

Locating region of interest for breast cancer in the mammographic image is a challenging problem in medical image processing. In this research work, we perform breast region segmentation followed by identification of the region of interest. Initially red, green and blue bands of the image are sliced into multiple segments based on intensity values and threshold. Masking operation is performed on each segment for understanding background and foreground(breast). Intersection of these segments would provide the breast segmentation. Next, we calculate the entropy of the segmented breast region. Here, quad division is performed based on the entropy. More entropy would result in additional quad division in the region of interest. This approach is tested on a DDSM datasets comprising of positive and negative mammographic images for breast location. Illustration is provided for the model. However, breast cancer detection needs radiologist’s attention in classifying normal and malignant cases. This leads to reduce false positive and negative rates when post processing is required for the detection of breast cancer.