Claim Missing Document
Check
Articles

Found 1 Documents
Search
Journal : Techne : Jurnal Ilmiah Elektroteknika

Klasifikasi Penyakit Pneumonia menggunakan Metode Convolutional Neural Network (CNN) Office, Irfan Handy; Arifuddin, Rahman; Hidayatulail, Basitha Febrinda
Techné : Jurnal Ilmiah Elektroteknika Vol. 23 No. 2 (2024)
Publisher : Fakultas Teknik Elektronika dan Komputer Universitas Kristen Satya Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31358/techne.v23i2.491

Abstract

Pneumonia is an acute lung infection that affects lung tissue. The disease can be caused by various pathogens such as viruses, bacteria, fungi, and others. COVID-19 pneumonia is a serious condition that requires special attention because of its contagious nature and its severe symptoms, including high fever, difficulty breathing, and lack of oxygen. Diagnosis usually depends on clinical symptoms and imaging techniques such as chest x-rays. With automatic classification technology, pneumonia detection becomes more efficient. The study used Convolutional Neural Network (CNN) with ResNet50 architecture to classify types of pneumonia, including viral pneumony and COVID-19, from chest x-rays. Research methods include literature reviews, data collection, pre-processing, modeling, training, testing, and evaluation using metrics such as accuracy, precision, recall, and F1 scores. Experiments with different epochs yield 99% accurate training data, 81% accurate validation data, and a lack of learning on models that influence accurability on validation.