Identification and Counting White Blood Cells and Red Blood Cells using Image Processing Case Study of Leukemia
This work addresses the time-consuming and costly manual blood cell counting in hematology clinics, though it is incremental as it applies existing image processing and fuzzy logic methods to a specific medical domain.
The study tackled the problem of manually counting blood cells for leukemia diagnosis by developing an image processing system to identify Acute Lymphocytic Leukemia (ALL) and Acute Myeloid Leukemia type M3 (AML M3) based on white blood cell morphology, achieving an accuracy of 83.65% in tests on 104 images.
Leukemia is diagnosed with complete blood counts which is by calculating all blood cells and compare the number of white blood cells (White Blood Cells / WBC) and red blood cells (Red Blood Cells / RBC). Information obtained from a complete blood count, has become a cornerstone in the hematology laboratory for diagnostic purposes and monitoring of hematological disorders. However, the traditional procedure for counting blood cells manually requires effort and a long time, therefore this method is one of the most expensive routine tests in laboratory hematology clinic. Solution for such kind of time consuming task and necessity of data tracability can be found in image processing techniques based on blood cell morphology . This study aims to identify Acute Lymphocytic Leukemia (ALL) and Acute Myeloid Leukemia type M3 (AML M3) using Fuzzy Rule Based System based on morphology of white blood cells. Characteristic parameters witch extractedare WBC Area, Nucleus and Granule Ratio of white blood cells. Image processing algorithms such as thresholding, Canny edge detection and color identification filters are used.Then for identification of ALL, AML M3 and Healthy cells used Fuzzy Rule Based System with Sugeno method. In the testing process used 104 images out of which 29 ALL - Positive, 50 AML M3 - Positive and 25 Healthy cells. Test results showed 83.65 % accuracy .