[–ß‚é]
|
Our final goal is to construct virtual violoncello instructor who can advice to make our skills better in the Biomechanical (Ergonomical) point of view. There are some layers in physical learning, which is based on neurological, kinetic, kinematic, and mental factors etc. As it is very hard to extract neurological and mental factors, we decided to extract kinetic and kinematic factors. For these 2 years we had some biomedical experiments with two kinds of biomedical sensors. One is Surface Electromyography (Surface EMG) which can measure how much muscle activates. The other is Retro Reflective Motion Capturing System which can track 3D motion of performers including X,Y,Z coordinates. In these experiments we found some circumstantial evidence of the importance of good spine posture during the performance. However, it is not so easy to analyze and interpret the biomedical time series data because it is very complicated and we have to consume many times to preprocess and analyze the data. I will show our basic idea to extract biomechanical knowledge by C4.5 from biomedical signals and experimental results by C4.5 in case that the data contains 2 experimental conditions and the time series data has some structures. I will also show you some ideas about these kind of analyses which I got during my short stay in this campus.
|
[ –ß‚é ]