Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Detection of Mycobacterium tuberculosis Using Real-time Detection Transformer

Rektorat UDINUS 2024 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2024
Goals
Record
7bbebb68-e352-4048-812d-4f14680fc366
Published
2026-10-30

Universitas Dian Nuswantoro (UDINUS) conducted an activity on the detection of Mycobacterium tuberculosis (MT) using a real-time detection transformer approach. Umam, Khairul; Purnama, I. Ketut Eddy; Rachmadi, Reza Fuad; and Wulandari, Sari Ayu participated in this activity. The activity addressed tuberculosis, described in the source as a severe disease caused by MT, and considered the use of sputum examination because this method is relatively cheap and easy. The source also states the importance of MT detection in medical personnel analysis.

The activity proposed MT detection through a one-stage object-detection approach. Two models were used: Detection Transformer (DETR) and Real-time Detection Transformer (RT-DETR), with different backbone variations. Experiments were also conducted using YOLOv5l and Faster-RCNN so that their evaluation results could be compared with those of DETR and RT-DETR. For DETR with the ResNet 101 backbone variation, the reported results were AP:50 = 0.681 and AP:50:95 = 0.276. RT-DETR with the HGNetv2l backbone variation obtained AP:50 = 0.837 and AP:50:95 = 0.147. YOLOv5l produced AP:50 = 0.75 and AP:50:95 = 0.326, while Faster-RCNN produced AP:50 = 0.696 and AP:50:95 = 0.30.

The source states that RT-DETR with the HGNetv2-l backbone demonstrated good performance in detecting MT. In relation to the stated goal, this activity represents health-focused research through the development and comparison of models for MT detection. The activity is institutionally framed as UDINUS research addressing a tuberculosis detection topic through a documented model evaluation.

Reference https://www.scopus.com/pages/publications/85208440993?origin=resultslist