The analysis of white blood cells in microscopy images allows the evaluation of hematic pathologies such as Acute Lymphoblastic Leukemia (ALL). Classification of white blood cells (WBC)is usually done manually by experienced hematologists. The efficiency and accuracy of this process depend on the skill and experience of the operator as well as his state of mind. On account of these reasons, the outcome of the classification may be undesirable.
In this paper, we present a methodology for fast automated segmentation of white blood cells from blood image sample. The focus lies in the classification algorithms viz. Random Forest and k Nearest Neighbor (kNN), which are used to classify cells as blast cells or not. The classification model is built from the features extracted from the blood smear images using the various image processing techniques.
http://www.tjprc.org/publishpapers/2-14-1530526922-1.IJCSEITRAUG20181.pdf
This repository contains the code for the GUI Application that was built to implement our research.