A Deep Learning–Driven Framework For Automated Segmentation Of Acute Leukemia In Peripheral Blood Smear Microscopic Images

Abstract

Early detection of acute lymphoblastic leukaemia symptoms can considerably improve patient's health condition. Detection or diagnosis of disease at the earlier stage is necessary for treating the patient. The microscopic image analysis requires subject experts and medical professionals. To meet the need, research on leukaemia has done by diverse scientist and the computational algorithms utilised for image analysis are highly influenced by the certain complexities. Occurrence of noise and blur edge portion necessitates diversified segmentation technique. Image segmentation is a way of breaking down an image information into several subdivisions that is determined as Image segments in order to reduce the image's complexities and make future processing or evaluation easier. This article research article focusses on the process of segmentation and the complexity is minimized by deep learning based convolutional neural network (CNN). The performance of the CNN is compared is compared with existing state-of-art technique whereby the CNN outperforms existing approaches.