Computer-aided Diagnosis Model for White Blood Cell Leukemia and Myeloma Classification using Deep Convolutional Neural Network

Preview this chapter:

Computer-aided Diagnosis Model for White Blood Cell Leukemia and Myeloma Classification using Deep Convolutional Neural Network, Page 1 of 1

< Previous page | Next page > /docserver/preview/fulltext/9789815274134/chapter-14-1.gif

Diagnosing white blood cell (leukocyte) diseases (Leukemia and Myeloma) is a thought-provoking task in the body. The abnormal growth of the leukocytes leads to an unbalanced immune system. Therefore, the automatic detection and classification of leukocytes will be the best aiding tool for the physician. This research work proposes a Computer-aided Diagnosis (CAD) model using the Deep Convolutional Neural Network (DCNN) to classify the white blood cell Acute Myeloid Leukemia (AML), Acute lymphoblastic leukemia (ALL), Myeloma, and its sub-types. The Gaussian distribution and k-means clustering segment the input image for future extraction. We utilized the Gray Level Covariance Matrix method to attain the texture features required to train the proposed DCNN model. The DCNN classifier is trained and tested with the mined features, and it detects the early stage of leukocyte cancer and achieves a classification accuracy of 97.8%. The precision, recall, and F1 score are achieved as 0.977, 80.955, and 0.966, respectively. We compared the performance of the proposed CAD model with the existing deep-learning classifier models. The analysis reveals that the proposed CAD model outperforms the existing methods.