With the economic support of Ministry of Science and Technology and several cooperative hospitals (including Far Eastern Memorial Hospital, Taoyuan General Hospital, Taichung Veterans General Hospital and Taichung Hospital) for several years. In addition, we have an international cooperation with City University of Hong Kong, Centre for System Informatics Engineering (CSIE), in order to develop the system of Fall Risk Assessment. The system have three features will promote the precision of Fall Risk Assessment System so far.


  1. With inclusion of the data of accelerometer and rotation angle
  2. With inclusion of the data of inertial sensor and image information
  3. With inclusion of machine learning and deep learning


The program gather the time series data of acceleration and rotation angle of go test with inertial sensor. Besides, we use the deep learning technique to learn the eigenvalue of fall risk of go test, including the deep learning of convolutional neural networks parameter setting, convolutional layer parameter design and train and test the different combination data, find out the most discriminability index of eigenvalue. We use the deep learning technique to learn the image information of go test.

3M TUG數據蒐集範例1
3M TUG數據蒐集範例2
3M TUG數據檔
3M TUG數據圖
AI 系統訓練