Database System for Storing Tuberculosis Sputum Sample Images as an AI Training Dataset

Abstract

The high prevalence of Tuberculosis (TB) in Indonesia puts Indonesia in the second-highest national TB prevalence in the world after India. This high prevalence can cause a failure to deliver medical treatments to TB patients, which is exacerbated by the disproportionate distribution of doctors in Indonesia. To address this issue, an AI system is necessary to help doctors in screening a large number of patients in a short time. However, to develop a robust AI for this purpose, we need a large dataset. This study aims to develop a database system for storing TB sputum sample images, which can be used as the dataset to train an AI for TB detection. The developed system can help doctors and health workers to manage the images during their daily job. After a period of time, the stored images can be utilized as the dataset to train AI.