This project aims to perform prediction of gestures captured by Myo, from the data generated by an I.A. predicts which gesture the user wishes to perform.
The database used to perform the predictions per second of each gesture will be the information generated by Myo instantly, where it will be aggregated per second.
The python used for development was 3.6.5. What things you need to install the software and how to install them
pandas==0.23.2
numpy==1.14.5
scikit-learn==0.19.1
pickle==0.7.4
myo-python==1.0.3
matplotlib==2.2.2
collections
time
To run this application you sould create a new virtualenv and install all the dependencies described below. If you prefer, you can use the requirements.txt included in this project to install all the dependencies on one single command.
pip install -r requirements.txt
You must install and connect the Myo device with its operating system, so this connection is important so that the myo-python package to see the device.
If you follow a sequence until you perform the predictions of the user's gestures, they are:
1. Calibration Process: For each user it is necessary to collect 30 seconds of each gesture. (Collect_data_final.py)
2. Preprocessing and modeling process: In this step, the data simulation process for the training will be performed and the 30 seconds collected will be used for testing. (Process_Model.py)
3. Gesture prediction: Here the prediction per second of each gesture performed by the user according to the model is already performed. (Predict_data_Myo.py)
The gestures defined were:
Ok - 👌
Cool - 👆
Fist - ✊
Open Hands - 🤚
PeaceandLove - ✌
Indicator - ☝
We are now on the development. As soon as we have a closed version, a tag should be created to hold the version 1.0 .
- Valter Eduardo - Initial version