Determination of blood components (WBCs, RBCs, and Platelets) count in microscopic images using image processing and analysis

摘要

Blood is one of the most essential parts of the human body, and it comprises of the RBCs, WBCs, and Platelets. Complete blood count characterizes the condition of well-being. Hence, segmentation and identification of blood cells is very important. Up to this day, many hospitals and health centers still use the old conventional method which involves manual counting of blood cells. This technique is time-consuming and prone to erroneous outcomes. On the other hand, there are some costly machines like Hematology Analyzer, which health centers cannot afford. This paper presents a raspberry-pi based image analysis system that is designed to segment and count blood cells from microscopic images of blood using Hue, Saturation, and Value (HSV) thresholding method and connected component labeling, respectively. Detection and counting of RBCs, WBCs, and Platelets have been done on ten microscopic images. Statistical analysis was performed to compare the values measured by the proposed system to the actual complete blood count test result of each patient. It shows that the proposed system has a 90% and up accuracy with respect to the actual CBC tests result. In addition, an android application was also developed to aid the user, especially those who are in rural areas, in assessing the number of blood cells, and to send the results from GUI to a doctor or specialist through short message service (SMS) for remote diagnosis.