UNAIR NEWS – Technology is advancing rapidly, including in image processing technology. The image processing technology can now be used in various fields, especially in the health sector.
However, if observed, the development of this technology has not been widely used in matters that are considered important. For example, in measuring white blood cell counts, because lacking or excessive white blood cell counts can be harmful to the body.
In Indonesia, maybe even in many countries, it is rare to find hospitals that use digital image processing technology as a method for calculating white blood cells. Digital white blood cell counts are considered to be more objective, accurate and effective than conventional methods.
In this regard, lecturer and researcher from Faculty of Science and Technology Universitas Airlangga (FST UNAIR) Franky Chandra Satria Arisgraha, S.T., M.T. created a leukocyte counting program using the Chain Code method, ” White Blood Cell Smart Counter ” or ” WBC Smart Counter “
Franky explained that the Chain Code method is a method in image processing that is useful for marking leukocytes as objects, as well as for knowing the patterns of existing leukocytes.
How it works
According to Franky, the research that has been running for a year is in the form of software used to observe photos of objects. The photo, he continued, from an image of an observation under a microscope and then captured. Then the file included in the big map file (the original file that has not been compressed, ed) is processed as input data with existing programs.
“In the form of software with inputs in the form of digital images captured visual field preparations observed using a microscope,” he explained.
The process conducted
Franky revealed that the study was an independent study with basic image processing. The research was started with extracting features by identifying the characteristics of the types of leukocytes based on morphological operations.
“So five types of leukocytes will be identified based on characteristics by means of morphological operations and feature extraction. We collect them based on the type of leukocytes, then we observe what features are appropriate so that they can be used as characteristics and differentiators for each type of leukocytes studied, “he said.
To avoid disruption in the identification process, Leukocyte images are separated from the background image through a segmentation process. The chain code is used in marking leukocyte images so that they can be grouped according to the characteristics of each type, after they are calculated.
“The composition and number of each type of leukocytes will be identified whether normal or not normal based on predetermined criteria,” he explained.
In the tests that have been done, in general, the program can function well in identification and calculation. But sometimes there are fail identifications because cell images appear to be overlapping, close together or attached, with similar shapes and colors so it can affect the accuracy of the calculation results.
“We are still trying to develop methods in order to find solutions to minimize or overcome misidentification or miscalculation,” he added.
With this research, Franky wanted to help labors who usually calculate them manually, so the calculation results are more objective, fast, and accurate. (*)
Author: Asthesia Dhea Cantika
Editor : Binti Q. Masruroh