Gesture-recognition-using-Myo-in-Tensorflow

• Use myo armband to collect multi-sEMG for every gesture by the Linux SDK: PyoConnect.
• Gesture recognition by DNN in Google's Tensorflow framework.
• The master branch is for the TensorFlow version 1.X(1.0-1.2), and there are another branch for the old version 0.X.
• Add 1D-CNN model as the gesture recohnition algrithm, and achieve online recognition.

linux:

ges_rec.py

• You can visualize the Neural Networks using tensorboard in linux after running ges_rec.py

tensorboard --logdir=/tmp/ges_rec_logs/

Windows:

conv1D_ges_rec.py -- model training by 1D-CNN

conv1D_ges_online.py -- online recognition by 1D-CNN

• the code is for windows using myo-python sdk
• You can visualize the Neural Networks using tensorboard in windows after running conv1D_ges_rec.py

tensorboard --logdir=F:/tensorflow_temp/ges_rec_logs/

requirement

Linux:

• PyoConnect
• Numpy
• Scikit-learn
• Tensorflow

Windows:

• myo-python
• Numpy
• Scikit-learn
• Tensorflow