Spatial Intuitionistic Fuzzy C-means with Calcifications enhancement based on Nonsubsampled Shearlet Transform to detect Masses and Microcalcifications from MLO Mammograms

Breast cancer is the most common form of cancer in women and the second leading cause of cancer death in this group. It is difficult, however, to diagnose cancer. Clusters of Microcalcifications and lumps are the two most prevalent indicators of breast cancer. MLO Mammograms can be processed digitally to find Masses and Microcalcifications. Most of the existing computer-aided techniques can detect either Masses or Microcalcifications, seldom both and no further measures are taken to mask the Pectoral muscle, even though its presence increases the probability of false positives. In this regard, a new work is presented to identify Masses and Microcalcifications in MLO Mammogram and specific steps applied to mask the Pectoral muscle. MLO Mammograms are analysed in this study and a technique is presented that will separate the process of identification of Masses and Microcalcifications. We separated the processing of Masses and Microcalcifications due to their size and characteristics. Microcalcifications are located by first enhancing the image with the nonsubsampled shearlet transform (NSST) and then using morphological operations to find bright spots. We also developed a unique image enhancement process and a modified spatial intuitionistic fuzzy C-means technique (SIFCM) to locate Masses. We applied this technique to the mini-MIAS database. The proposed approach has shown promising results.