UNITED: Utilizing AI for Non-Invasive Early Detection of Diabetes Mellitus through Tongue Analysis

Ameera Indriani Lubis, Azkia Najla Shazia, Fashila Firdausi Zahra, Mahira Alyna Nisaa, Ratu Dinara Nafeeza Nurdin

1) Ameera Indriani Lubis, MAN 4 Jakarta, South Jakarta, Indonesia (ameera.indriani.1357@gmail.com)

2) Azkia Najla Shazia, MAN 4 Jakarta, South Jakarta, Indonesia (iphoneazkia@gmail.com)

3) Fashila Firdausi Zahra, MAN 4 Jakarta, South Jakarta, Indonesia (fashila.zahra04@gmail.com)

4) Mahira Alyna Nisaa, MAN 4 Jakarta, South Jakarta, Indonesia (mahira.nisaa12@gmail.com)

5) Ratu Dinara Nafeeza Nurdin, MAN 4 Jakarta, South Jakarta (aradinars31@gmail.com)

Abstract

Diabetes mellitus (DM) is a chronic, non-infectious disease that currently affects around 540 million people globally, with projections suggesting this number could reach 783 million by 2045. In Indonesia, the prevalence is notably high, with 10.8% of adults impacted, making the country the second- highest for diabetes-related mortality. Early detection and management are crucial in addressing the growing burden of DM. While traditional non- invasive methods like tongue diagnosis have been used in various cultures, these approaches often face limitations due to their subjective nature and dependency on practitioner expertise, leading to inconsistent results. The UNITED (Utilizing AI for Non-Invasive Early Detection of Diabetes Mellitus Through Tongue Analysis) project aims to tackle these challenges by integrating Huskylens, an AI-enabled vision sensor, to perform real-time, non-invasive tongue image analysis. By employing machine learning algorithms, this tool provides a standardized and objective method for early DM detection, improving diagnostic accuracy and capturing subtle symptoms that might be missed by traditional methods. This innovation not only promotes early intervention but also supports SDG 3 (Good Health and Well-being) by enhancing healthcare accessibility and efficiency, offering a reliable, user-friendly alternative to conventional blood glucose tests.

Keywords : Diabetes Mellitus, Early Detection, Tongue Diagnosis, Huskylens, Non-Invasive Detection Tool.