Multiple Myeloma Prediction from Bone-Marrow Blood Cell images using Machine Learning

Myeloma is a type of blood cancer that affects the plasma cells in the bone marrow, which produces antibodies that help the immune system to fight against outside aggression. Myeloma results in production of abnormal antibodies thereby weaking the function of the immune system. Out of the different types of blood cancers, the proposed research provides a robust mechanism for Multiple Myeloma (MM) prediction using 85 Microscopic blood images that were captured from bone marrow aspiration of patients suffering from the disease. The proposed work eradicates the probability of errors in the manual process of feature extraction by employing Convolutional Neural Network (CNN) for it and this is followed by training the model with Artificial Neural Networks (ANN), Support Vector Machine (SVM), Random Forest algorithms for classification. Along with SVM and Random Forest we used Random Search Optimizer for finding the suitable set of hyper parameters for better results. The overall accuracy was recorded to be 100%, for CNN-ANN model. Thus, the model can be used effectively for determining the Multiple Myeloma from the cell images.