2020 Volume 41 Pages 217-225
The purpose of this study was to test if lung volume during swimming can be estimated from optical motion capture system. To achieve this goal, we tried to develop a digital human model expressing respiratory motion. A male swimmer performed lung function tests during the quiet standing and tethered front crawl swimming under the water, and respiratory flow rate and three-dimensional （3D） motion data were measured synchronously. Respiratory flow rate was integrated to obtain lung volume （measured lung volume） as a function of time. 3D motion data of trunk segment was collected by an underwater motion capture system. 3D shape data of the whole body was scanned for the swimmer on dryland. Digital human model expressing respiratory motion was created based on the 3D shape data. Then, digital human model was fitted to 3D motion data by using inverse kinematics computation, and lung volume was estimated from the change in volume of the model （estimated lung volume）．Measured and estimated lung volume were compared by using the Intra-class correlation coefficient （ICC）．As a result, ICCs during standing were 0.981 for quiet breathing, 0.995 for force vital capacity, and 0.972 for maximal voluntary ventilation （P < 0.001）．On the other hand, ICC was smaller during swimming compared with values during standing, with being 0.717 （P < 0.001）．These results suggest that our digital human model with motion capture data under the water can accurately estimate lung volume at least in lung function tests during standing. Further study is needed to improve our digital human model to estimate lung volume more accuracy during swimming.
