Medical Image Analysis of Multiple Myeloma Diagnosis Using CNN and KNN Based Approach

Leukocytes, which are produced in the bone marrow, account for about 1% of all blood cells. When the production of these neutrophils becomes unchecked, cancer of the blood occurs. In order to better classify Acute Lymphoblastic Leukemia (ALL) and Multiple Myeloma (MM), two of the three types of cancer studied, the proposed study makes use of the SN-AM dataset. Acute lymphoblastic leukemia (ALL) is a cancer characterized by an excess generation of lymphocytes in the bone marrow. However, malignant cells from multiple myeloma (MM) do not enter the bloodstream and instead remain in the bone marrow. Healthy stem cells are crowd out, which prevents the body from producing enough new blood cells. In the past, this would necessitate the dedicated time and energy of an expert. The proposed approach makes use of preprocessing, segmentation, feature extraction, and evaluating model performance. Median filters are utilized in the preparation of data because the capturing and transmitting of medical images is often hindered by noise. Cell Segmentation, as well as Nucleus and Cytoplasm Segmentation, are two common techniques used in segmentation. Gaussian pyramids is used for feature extraction. The CNN-KNN Model is used to evaluate the performance of the model. The KNN and CNN models are used as benchmarks against which the suggested method is assessed. It far surpasses simpler methods in comparison.