複数姿勢からの関節位置推定による関節可動域計測システム

抄録

Measuring the range of motion (ROM) of the human body is essential for the quantitative evaluation of rehabilitation for patients with cerebrovascular disease. However, manual measurement of ROM requires enough knowledge and experience and is laborious work. Therefore, the computer-assisted system helps reduce such situations. Kinect sensor can automatically detect many joint positions in the human body and is considered adequate for applying the above system. The purpose of this study was to develop an automatic ROM measurement system by using a Kinect sensor. The system is aimed at detecting the motion for putting on and taking off clothes. For accurate ROM measurement in the motion, some joint positions, which a Kinect sensor did not give those, were estimated from the data of multiple postures. Then, ROMs on the thoracic and lumbar spine were calculated under flexion, extension, lateral bending, and rotation, and on the shoulder girdle under flexion and extension. The developed system was applied to ROM measurements for four healthy adults, and its result was compared with the results of an expert occupational therapist. Most of the measurement error between the system and an expert maintained less than 5 degrees. This system helps reduce the therapist burden in the assessing thoracic and lumber spine and shoulder girdle mobility and provides feedback to the patients on the effectiveness of rehabilitation. Further development of the system is expected to improve the current rehabilitation situation.