Universitas Indonesia (UI) through the Medical Physics & Biophysics Field Group (KBI) and KBI
Physics Instrumentation – The Physics Department of the Faculty of Mathematics and Natural Sciences, University of Indonesia (FMIPA UI) has developed DSS-CovIDNet, a tool for predicting COVID-19 pneumonia cases using a deep-learning artificial intelligence-based program. The program was designed by a team of master’s students and alumni from the Physics Department, FMIPA UI, who are members of the AIRA (artificial intelligence for radiological applications) research team under the direction of Prof. Dr. Djarwani S. Soejoko, FIOMP.
and Prawito, Ph.D.
DSS-CovIDNet using the concept of convolutional neural network (CNN) to classify chest X-ray images into 3 (three) groups, namely COVID-19 pneumonia, Non-COVID-19 pneumonia, dan paru normal dengan akurasi mencapai 98,44%. Koordinator Tim AIRA, Lukmanda Evan Lubis S.Si., M.Si. He stated, “DSS-CovIDNet is expected to contribute to increasing diagnostic confidence and reducing the burden on radiologists with their high workload related to diagnosing and monitoring COVID-19 cases.”
UI Vice Chancellor for Research and Innovation Prof. Dr.rer.nat. Abdul Haris, M.Sc He added, “The high level of accuracy makes this tool superior. We’re also opening up data access in the hope of making it easier for researchers to help refine this program.”
Research related to pneumonia detection is not only carried out by this one research group, but there are three other interdisciplinary research teams at UI to detect COVID19 pneumonia using artificial intelligence based on radiological data. The researchers are the UI Faculty of Medicine (FKUI) Research Team in collaboration with DELFT Imaging CAD4COVID under the direction of Dr. Eric Daniel Tenda, SpPD, and Dr. Benny Zulkarnaien, SpRad(K). In addition to them, there is a FKUI research group under the direction of Dr. Cleopas Martin Rumende, Sp.PD-KP and Dr. Telly Kamelia, SpPD., KP to develop a detection algorithm. Next, a research team from the UI Faculty of Computer Science (Fasilkom UI) led by Mirna Adriani, Dra., B.Sc., Ph.D., Dina Chahyati, S.Kom., M.Kom. who collaborated
with a team from the AI Center Fasilkom UI.
A description of the method and preliminary results using open source datasets can be downloaded at https://arxiv.org/abs/2005.04562, while the validation process using anonymous Indonesian patient data has been initiated in collaboration with the staff of the Department of Radiology, Faculty of Medicine, University of Indonesia (FKUI), the Radiology Unit team at UI Hospital (RSUI), and the Radiology Installation of Cibinong Regional General Hospital.
The program’s development is fully supported by the Faculty of Mathematics and Natural Sciences, University of Indonesia (UI). It can be accessed for testing at http://sci.ui.ac.id/detectcovid/ using an access key, which can be requested free of charge via email to aira.medphy.ui@gmail.com. Users are limited to medical and healthcare personnel in the field of radiology.
