https://github.com/sohhamseal/what-are-you-pointing-at

Raw Surface-EMG signal classification using DL models
https://github.com/sohhamseal/what-are-you-pointing-at

alexnet cnn emg-signal fourier-transform myo-armband rccgnet resnet-18 resnet-34 signal-processing sliding-window vgg16 zfnet

Last synced: 2 months ago
JSON representation

Raw Surface-EMG signal classification using DL models

• Host: GitHub
• URL: https://github.com/sohhamseal/what-are-you-pointing-at
• Owner: SohhamSeal
• Created: 2024-03-29T14:39:21.000Z (10 months ago)
• Default Branch: main
• Last Pushed: 2024-09-11T03:29:10.000Z (4 months ago)
• Last Synced: 2024-09-11T08:37:55.173Z (4 months ago)
• Topics: alexnet, cnn, emg-signal, fourier-transform, myo-armband, rccgnet, resnet-18, resnet-34, signal-processing, sliding-window, vgg16, zfnet
• Language: Jupyter Notebook
• Homepage:
• Size: 27.3 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# What are you pointing at ??

## EMG Signal Gesture Recognition

This repository contains code and resources for conducting gesture recognition using EMG (Electromyography) signal data. The dataset used in this project can be found on Kaggle [here](https://www.kaggle.com/datasets/sojanprajapati/emg-signal-for-gesture-recognition).



## About the Dataset

The dataset consists of raw EMG data recorded from individuals wearing a MYO Thalmic bracelet on their forearm while performing various hand gestures. Each gesture is labeled with a specific category, and the dataset includes information such as time stamps and channel readings from the sensors on the bracelet.

## Work Conducted

In this project, we conducted the following steps to preprocess the data and classify the gestures:

1. **Data Preprocessing**:

   - Cleaned the dataset to handle any missing or erroneous data points.

   - Applied sliding window Fourier transform to segment the EMG signals into smaller windows, which helps capture temporal patterns in the data.

   - Extracted features from the Fourier transformed signals to reduce dimensionality and focus on relevant information for classification.

2. **Model Selection**:

   - Experimented with various deep learning models suitable for time-series data classification, including:

     - AlexNet

     - ResNet18

     - Resnet34

     - VGGNet16

     - ZFNet

     - RCCGNET

   - Chose models with architectures capable of handling 1D input data while maintaining the convolutional and pooling layers for feature extraction.

3. **Model Training and Evaluation**:

   - Split the preprocessed data into training and testing sets uisng k-fold Cross Validation.

   - Trained each selected model on the training data using appropriate loss functions and optimization techniques.

   - Evaluated the trained models on the test data to assess their performance in classifying different gesture categories.

   - Utilized metrics such as accuracy, precision, recall and F1-score to measure the classification performance of each model.