Size-based Feature Extraction on Blood Cells Calculation Process Using k-Means Clustering

Abstract

In human blood, there is a blood plasma that has the most important function in the human body. Blood Plasma consists of red blood cells (erythrocytes), white blood cells (leukocytes), and platelets. Of these three blood cells can be known health condition in humans, by counting the number of blood cells contained in the blood. In this modern era, computing systems are very helpful for various areas, one in the medical field to calculate and classify the shape and size of cells in the blood. Therefore, this research creates a system that is capable of extraction of features on blood cells using morphological methods as well as grouping of multiple blood cells based on size using the method K-Means Clustering to calculate the number of Blood cells. The image processing patterns in blood cells begins with the image capture of blood cells using a microscope, the image of blood cells in the extraction based on size and shape using morphological surgery, then from the process of extraction of blood cells will be in Group into three based on the size and shape of the erythrocytes, leukocytes, and platelets using K-Means Clustering.The processing of image patterns in blood cells begins with the image capture of blood cells using a microscope, the image of blood cells in the extraction based on size and shape using morphological surgery, the results of the extraction process of blood cells will be grouped into three based on size and shape using K-Means Clustering. The grouping process will be known to the number of each of these blood cells. The method used in this study resulted in the percentage of successful object recognition and a very good calculation. Results of the percentage generated 60% success and error 40%, it is still a lot of improvement with the process of object recognition based on the characteristics of the blood cells using other methods.