Open AccessDigital Image Segmentation Resulting from X-Rays of Covid Patients using K-Means and Extraction Features Method
Author(s) -
Dhian Satria Yudha Kartika,
Anita Wulansari,
Hendra Maulana,
Eristya Maya Safitri,
Faisal Muttaqin
Publication year - 2021
Publication title -
international journal of computer, network security and information system (ijconsist)
Language(s) - English
Resource type - Journals
ISSN - 2686-3480
DOI - 10.33005/ijconsist.v3i1.55
Subject(s) - confusion matrix , segmentation , covid-19 , computer science , artificial intelligence , pattern recognition (psychology) , stage (stratigraphy) , process (computing) , pixel , feature extraction , digital image , feature (linguistics) , image processing , medicine , image (mathematics) , disease , pathology , paleontology , linguistics , philosophy , infectious disease (medical specialty) , biology , operating system
The COVID-19 pandemic has significant impact on people's lives such as economic, social, psychological and health conditions. The health sector, which is spearheading the handling of the outbreak, has conducted a lot of research and trials related to COVID-19. Coughing is a common symptoms among humans affected by COVID-19 in earlier stage. The first step when a patient shows symptoms of COVID-19 was to conduct a chest x-ray imaging. The chest x-rayss can be used as a digital image dataset for analysing the spread of the virus that enters the lungs or respiratory tract. In this study, 864 x-rays were used as datasets. The images were still raw, taken directly from Covid-19 patients, so there were still a lot of noise. The process to remove unnecessary images would be carried out in the pre-processing stage. The images used as datasets were not mixed with the background which can reduce the value at the next stage. All datasets were made to have a uniform size and pixels to obtain a standard quality and size in order to support the next stage, namely segmentation. The segmentation stage of the x-ray datasets of Covid-19 patients was carried out using the k-means method and feature extraction. The Confusion Matrix method used as testing process. The accuracy value was 78.5%. The results of this testing process were 78.5% of precision value, 78% of recall and 79% for f-measure