This page contains information about the screening software and program in the paper "Cervical myelopathy screening with machine learning algorithm focusing on finger motion using non-contact sensor".
- Leap Motion (Leap Motion, San Francisco, CA, USA) ver2.1
- Unity
- Python 3.6
You can execute our screening system easily if you can make new data.
Install or copy "CM_screening-main" from the page top.
We prepared a svm learning model "svm_data.pkl", which is using a combination of parameters (the position and direction).
- Prepare a computer with a macOS (Apple Inc., Cupertino, CA, USA) and Leap Motion.
- Connect Leap Motion to the computer.
- Uncompress the compressed file "HANZM_exe".
- Open the folder "HANZM_exe" and start HANZM ver2.1.
- Decide and input the 8-digit ID ******** every each person.
- Perform the examination by gripping and releasing fingers 20 times on the Leap Motion according to the Measurements in the paper.
- Get the examination data at "HANZM_exe/HANZM ver2.1/Contents/Log".
- Add the examination data from Unity into "classification-data/********"(8-digit ID).
Please make a directory every user, which is named as "********"(8-digit ID). - Execute cm_classification.py. After it, the classification result is shown.
Our data from Leap Motion has 35 features.
- speed : 0~4
- position : 5~19
- direction : 20~34
You can make your svm learning model for classification.
Please get some people's data by a similar flow .
- Add the examination data from Unity into "10s-test-data/Patient/********" or "10s-test-data/Health/********"(8-digit ID).
Please make a directory every user, which is named as "********"(8-digit ID).
If the data is CM patient's one, add into the former. If else, add into the latter.
- Execute make_model.py. After it, the classification model is made, and the result of the Leave-one-out evaluation is shown.
