부산대학교 도서관

To predict the risk of pressure injuries in home‐care, it is important to quantify deformation of internal tissue. However, it is difficult for users in home‐care to use conventional modalities such as MRI because the modalities are expensive and not portable for the trunk. Therefore, we have developed a deformation estimation method in a sitting position with the use of machine learning and a low‐cost portable pressure mapping system. We defined the deformation as estimated soft tissue thickness values on each cell of pressure distribution. The thickness on a cell was estimated from measured pressure distribution and no‐load thickness. A no‐load thickness was prepared before estimation using a model. In this article, the estimation method was applied to two single‐layered phantoms which were constructed from wood as an ischial tuberosity and two gels as muscle and fat. The estimation errors were only 1.03% and 2.56% in the muscle and fat phantoms, respectively. We confirmed the estimation errors were adequately small values. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]