Due to technological advancements, electromyography (EMG) is now used for more than only diagnosis, with potential applications including movement analysis and the precise and flexible control of manipulators in the field of aided medicine. Intuitive and precise recognition of multiple motions is possible with a powerful classifier using surface electromyography-based gesture recognition systems. The Mayo armband is a Bluetooth-enabled, low-power wireless sensor that generates a reliable EMG reading. The Myo armband detects and records upper-limb electrical muscle activity. Artificial intelligence and deep learning-based models generally adopted with excellent outcomes in many fields. In this paper, an artificial neural network-based method for EMG gesture prediction using pre-trained DNN characteristics is proposed. The proposed method analyses the CapgMyo standard benchmark dataset, which maps eight classes of hand movement recognition to data collected from participants via the Myo wristband EMG signals. The results demonstrate that the suggested classification method, which employs an artificial neural network classification model with deep features of EMG signals, beats the other current methods by a significant margin with accuracy of 94.8%.

DNN, DWT, Electromyography, Hands Gesture Recognition, Machine Learning