Estimation of tidal volume in a pediatric setting using depth camera

Abstract

Respiratory failure is one of the diseases of humans, regardless of age. It can occur at any time and can be detected by the study of respiratory measurements such as inhaled air volume and breathing frequency by the clinician. But the analysis of these parameters is very difficult with the naked eye, especially when it comes to children. To overcome this problem encountered during diagnosis, this research project aims to develop a more efficient and effective approach to the analysis of respiratory measurements. Mainly, the design and development of a computerized and autonomous system based on 3D images, which would allow the study of the volume of inspired and expired air based on the thoracoabdominal movement as well as the study of the respiratory frequency to detect a respiratory insufficiency. A methodology has been proposed to achieve this in the pediatric department. This methodology consists of the acquisition of 3D images of the patient during breathing with a KINECT Azure camera positioned above the patient, followed by a reconstruction of the image obtained and then the analysis of this image by creating sinusoidal graphs to represent the respiratory cycles in the acquisition using Matlab and CloudCompare. The detection of the respiratory cycles makes it possible to calculate by subtraction the volume of inspired air. The respiratory cycles are detected thanks to the position of the abdomen and the thorax at each instant (lowered in exhale and raised in inhale). The reliability and accuracy of the method has been examined by performing several tests on real patients and mannequins, comparing the results obtained by the method with the real results. The end of inhale and exhale tests were compared to the results detected visually by a physician, resulting in errors of 1-3 frames. The tests were performed on a mannequin and on a real patient. A real patient test gave frame 1 for end of inhale first cycle, frame 27 for second cycle and frame 60 for third cycle. For the end of exhale, we got the frames 13, 43 and 70 for first, second and third cycle respectively. After the analysis, the clinician detected frames 3, 27, 57 for end of inhale of the first, second and third cycle respectively. For the end of exhale the frames 12, 45 and 70 for first, second and third cycle. The tests were compared to ventilator values for respiratory volume, resulting in an average error of 14%. The tests show the effectiveness of the system and prove that it is a promising system. It will be of great help in situations of lack of experts, pandemic and especially in pediatrics where the patients are young with very small respiratory volumes and fragile immune system.

Date	12 Sept 2023
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Rita Noumeir (Supervisor) & Philippe Jouvet (Co-supervisor)