Hybrid Deep Learning Models for Hand Gesture Recognition with EMG Signals

Hand Gesture detection utilizing Electromyography (EMG) signals is a dynamic field of study with a wide range of applications, including human-machine interfaces and prosthetic devices. In recent years, several deep-learning algorithms have been created to classify EMG data. This paper examines the impact of hybrid models, which integrate CNN and RNN models, on hand gesture categorization using EMG signals. A comparative analysis is conducted on the performance of three hybrid models: CNN-GRU, CNN-LSTM, and CNN-BiLSTM. The performance assessment is conducted on the Ninapro DB1 dataset. The findings demonstrate that CNN-BiLSTM models outperform the other three models in terms of classification accuracy, achieving a superior result of 79.38%. But when the complexity of the model is a concern, CNN-GRU models may be a better choice.