Related Experiment Video
Updated: Jul 1, 2025
08:15
Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision
Published on: March 28, 2025
480
Hand gesture recognition with deep residual network using Semg signal
Abid Saeed Khattak1,2, Azlan Bin Mohd Zain1, Rohayanti Binti Hassan1
1Faculty of Computing, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia.
Biomedizinische Technik. Biomedical Engineering
|March 8, 2024
Summary
This study introduces a new hand gesture recognition classifier, the Sewing Driving Training based Optimization-Deep Residual Network (SDTO_DRN). The SDTO_DRN model demonstrates high accuracy in recognizing hand gestures from sEMG signals.
Area of Science:
- Biomedical Engineering
- Machine Learning
- Human-Computer Interaction
Background:
- Surface Electromyography (sEMG) signals from forearm muscles contain valuable data for decoding hand movements.
- Accurate hand gesture recognition is crucial for advanced human-computer interaction and prosthetics.
Purpose of the Study:
- To design and develop an effective classifier for hand gesture recognition using sEMG signals.
- To introduce the Sewing Driving Training based Optimization-Deep Residual Network (SDTO_DRN) model.
Main Methods:
- Utilized sEMG signals captured from forearm muscles.
- Applied Gaussian filtering for signal pre-processing and feature extraction.
- Employed the SDTO_DRN model, integrating Sewing Training Based Optimization (STBO) and Driving Training Based Optimization (DTBO) for feature selection and network fine-tuning.
- Validated the model on the MyoUP and putEMG datasets.
Main Results:
- The SDTO_DRN model achieved a maximum accuracy of 0.943.
- Key performance metrics included a True Positive Rate (TPR) of 0.929, True Negative Rate (TNR) of 0.919, Positive Predictive Value (PPV) of 0.924, and Negative Predictive Value (NPV) of 0.924.
- The results indicate superior performance in hand gesture recognition.
Conclusions:
- The proposed SDTO_DRN model provides accurate and effective hand gesture recognition.
- This advancement holds potential for improving human-computer interaction systems.
