Open Conference Systems, MISEIC 2019

Font Size: 
Application of Convolutional Neural Networks Method in Predicting Pneumonia Diseases Based on X-Ray Test Results
Rivan Pratama Yuda, Zuherman Rustam, Chelvian Aroef, Hamimah Alatas

Last modified: 2020-02-02

Abstract


Pneumonia is an infectious disease that attacks the lungs so that the air sacs in the lungs are experiencing inflammation and swelling. Pneumonia is usually often known as wet lungs. Pneumonia can attack all walks of life, both adult as well as toddlers. From the phrase Dr. Christina Widaningrum as the head of Sub Directorate (Kasubdit) of Acute Respiratory tract infections (RESPIRATORY) Ministry of health (Kemenkes), Indonesia was ranked 10th in the world in the case of death from pneumonia. To find out the condition of a healthy lung  or not from someone patient needs to do an x-ray. However, based on the results of the x-ray lung conditions or exposed to pneumonia by bacteria or viruses can we know based on the explanation from doctors in their field. Therefore, this study is intended to help doctors and patients so that knowing more clearly and more confident again that picture of lung x-ray results included into the category of a healthy lung or exposed to pneumonia by bacteria or viruses. To know it, can be assisted with the use of method or algorithm of Convolutional Neural Networks (CNNs). Why should we use CNNs method? Because CNNs is a deep neural network that’s suitable for image data. And the other side, CNNs has a good accuracy value for classification image to predict something. Later, we use the image data from the patient’s lung x-ray results so that the CNNs will work by finding an image that corresponds to the category of the lungs of someone who belongs into the lungs are healthy or are exposed to pneumonia because bacteria or viruses. And based on the score showed that the performance of CNNs attained an accuracy rate of 97%.

Keywords


Pneumonia; Convolutional Neural Networks; X-Ray

Full Text: PDF