Identification of Bacilli Bacteria in Acute Respiratory Infection (ARI) using Learning Vector Quantization

Abstract

Two diseases that include Acute Respiratory Infections (ARI) are diphtheria and tuberculosis. Both diseases have a large number of sufferers and can cause extraordinary events (KLB). One of the achievement indicators of infectious disease control and management programs is discovery. However, the limited number of medical analysts causes the discovery process (examination) long and subjective. To help with this problem, a bacillus identification system was created for early detection of Acute Respiratory Infections (ARI). This system is an implementation of computer vision. The data used are preparations of the bacteria Mycobacterium tuberculosis and Corynebacterium diphtheriae obtained at Besar Laboratorium Kesehatan (BBLK) Surabaya. The parameters used are the area, perimeter and shape factor. The Learning Vector Quantization (LVQ) method can classify and identify bacillus bacteria that cause acute respiratory infections with a training accuracy of 97% and a test accuracy of 86% with a learning rate of 0.01 and a reduced learning rate of 0.25.