Towards Efficient Segmentation and Classification of White Blood Cell Cancer Using Deep Learning

Abstract

White Blood cell cancer is a plasma cell cancer that starts in the bone marrow and leads to the formation of abnormal plasma cells. Medical examiners must be exceedingly selective when diagnosing myeloma cells. Moreover, because the final judgment is dependent on human perception and judgment, there is a chance that the conclusion may be incorrect. This study is noteworthy because it creates a software-assisted way for recognizing and identifying myeloma cells in bone marrow scans. MASK-Recurrent Convolutional Neural Network has been utilized for recognition, while Efficient Net B3 has been used for detection. The mean Average Precision (mAP) of MASK-RCNN is 93%, whereas Efficient Net B3 is 95% accurate. According to the findings of this study, the Mask-RCNN model can identify multiple myeloma, and Efficient Net B3 can distinguish between myeloma and non-myeloma cells.