A Non-Contact Height Measurement Method Using Mediapipe and Yolov8 in a Three-Dimension Space

Phu Nguyen Trung (Hanoi University of Industry)
Dat Dong Quoc (Hanoi University of Science and Technology)
Kien Nguyen Phan (Hanoi University of Science and Technology)

Journal of Applied Engineering and Technological Science (JAETS)
Vol. 7 No. 2 (2026): Journal of Applied Engineering and Technological Science (JAETS)

Accurate human height measurement is an essential clinical parameter for calculating body mass index and assessing patient health status. However, traditional contact-based or manual methods are impractical for patients with limited mobility or during emergency conditions. This study addresses this limitation by developing a non-contact 3D height estimation system using a smartphone-based vision approach. The system employs MediaPipe to extract skeletal landmarks and a regression model to predict human height from bone segment lengths and a reference object. Experimental evaluation on 166 subjects across multiple positions demonstrates a high accuracy with an average error rate of 1.54 ± 0.64%, outperforming most existing camera-based systems. The proposed method offers a practical, low-cost, and portable solution for medical height assessment, particularly beneficial in clinical and emergency settings. Potential limitations include dependence on camera angle and lighting conditions, which will be addressed in future work.