/sEMG-Neural-Net

Neural network for classifying electromyographic signals into distinct gestures. Additionally, a comparison of CNN vs LSTM implementations.

Primary LanguageJupyter Notebook

sEMG Signal Neural Network Classifier

Electromyography is the study of electrical signals from skeletal muscles. These signals (sEMG signals) differ based on muscle movement and can be captured with specialized peripherals. This project is an attempt to construct and train a neural network to interpret the raw composite signal data, and classify the sequence data as one of the following grips:

sEMG_Basic_Hand_movements_upatras/grasps_en.PNG

Languages and frameworks used:

Models were deployed to Paperspace for training and hyperparameter tuning.