Determination of Blood Components Count in Real Time Microscopic Images Using Image Processing

Abstract

RBCs, WBCs, and platelets are all components of blood, which is one of the body's most vital organs. The state of health is characterized by a complete blood count. Consequently, segmenting and identifying blood cells is crucial. Many hospitals and healthcare facilities still count blood cells manually using the outdated conventional approach. This method takes a lot of time and can lead to inaccurate results. On the other hand, health centers cannot afford expensive equipment like the hematology analyzer. This study describes a raspberry-pi based image analytic system that is meant for the segmentation and counting of blood cells from microscopic blood-smear images using linked component labeling and the Hue, Saturation, and Value (HSV) thresholding method, respectively. On 10 microscopic pictures, RBCs, WBCs, and platelets have been identified and counted. A statistic-based analysis was done to contrast the values recorded by the suggested system with each patient's actual total blood count test result. It demonstrates that the proposed approach has an accuracy of 90% or higher in relation to the real results of CBC testing. Additionally, an Android app was created to assist users, particularly those in the non-urban regions, in determining the amount of blood cells and to send the results from the Graphical User Interface to a certified physician or expert through Short Message Service for remote diagnostic purposes [3].