The Non-Contact Height Measurement Method Using MediaPipe and OpenCV in a 2D Space

Authors:
Publisher website
Google Scholar
Abstract
Body height is a fundamental parameter in healthcare as it allows for the calculation of various treatment-related metrics. However, for critically ill patients in intensive care units (ICUs), having them move or stand straight for height measurement is not feasible. Therefore, the research is aimed at developing a non-contact method for determining the height of critically ill patients. However, in the first phase, research focuses on developing a non-contact height measurement method in a 2D space to evaluate the measurement accuracy before targeting critically ill patients specifically. The research utilizes a video camera and the MediaPipe library to extract the coordinates of the human body's skeletal frames. These coordinates, representing the X and Y coordinates in a 2D space, are used to calculate the lengths of the body's skeletal frames and, consequently, to determine the height of the individual. Additionally, the study employs a data training model to improve measurement accuracy. Based on the calculated formulas and the execution of the training model on a dataset of 43 individuals, with 36 sets for training and 7 sets for testing, the measurement results of the system show an error ranging from Ocm to 1.6cm with a height range from 163 cm to 175.5 cm. The obtained level of accuracy indicates that this margin of error is acceptable for critically ill patients. In the future, the research will be expanded to determine the height of critically ill patients in a 3D space, considering different lying positions, with the aim of developing the most effective application software for critically ill ICU patients.