患者の動向監視支援システムの開発

―デプスセンサの実装と評価―

抄録

In recent years, Japan has become a hyper-aged society, and accidents caused by falls and stumbles of dementia patients have been increasing. In previous research, a motion monitoring support system using depth sensors has been developed to reduce these accidents. However, among the five available depth sensors, Kinect2 and RealSense D455/L515 are no longer in production and are difficult to obtain. Therefore, we introduced a new depth sensor, TEDTOF manufactured by Tokyo Electron Device, and implemented it in the existing system. Two image filters and one image processing were used to reduce noise in the distance image obtained from the TEDTOF. In addition, a comparison experiment was conducted to discriminate the patient's condition among the five depth sensors used in this study. The results showed that TEDTOF had the best recognition accuracy and RealSense L515 had the lowest recognition accuracy.