Segmentation Techniques for Multiple Myeloma Using Machine Learning

The worldwide decrease in life expectancy rate distracts the kind concern of humans towards health issues. Cancer is a very common disease, and among cancers, the possibility that it is Multiple Myeloma (MM), a cancer of Plasma Cells in which a group of plasma cells becomes cancerous and multiplies, is very high because it’s 3 rd most commonly occurring form of cancer. But, like other cancers it’s very difficult to detect at an early stage. The disease can damage the bones, immune system, kidneys and red blood cell count. It is now essential for technology to advance and build a reliable working system which provides a exact accuracy to detect this type of cancer and benefit humankind. The development of a computer aided diagnosis system is presented for Plasma cell segmentation in case of Multiple Myeloma, based on image processing and machine learning. The stain normalized microscopic images of the bone marrow of patients were analyzed by applying various image processing as well as machine learning algorithms (such as k-means and fuzzy c means) to figure out the plasma cell segmentation among the patients.