An Effective Recognition of Pointer Movement By Means of Hilbert Huang Transform and Deep Neural Network Algorithm

Electromyography (EMG) is a common signal used to detect and categorize various sorts of bodily motions, including, but not limited to, varied positions, gestures, and actions. Because of its efficiency and ease of use, the surface Electromyogram (sEMG) signal was mostly used for analysis. Due to its shortened calculation time, higher efficiency, and good accuracy, the EMG signal was created from the neural excitations of the muscles using an EMG driven neuromuscular model. Time-frequency domain parameters derived from electromyography were used to discuss and quantify musculoskeletal activity. The Hidden Markov Model (HMM) was used to specify the musculoskeletal moment process. Methods based on the Hilbert-Huang Transform (HHT) are used to get around the WT's shortcomings. Artificial neural network (ANN) based methods are utilized for gesture recognition automation. In order to categorize the hand movements, we create a collection of Feed Forward Neural Networks (FFNN) that are educated using a back propagation approach. However, the network has two major drawbacks: it takes a long time to train, and it can't spot outliers. This is why hand gesture recognition utilizes Cascaded Feed Forward Neural Networks. Providing sufficient hidden neurons, the cascaded networks can learn any input-output relationship in a shorter amount of time. It can recognize unusual cases throughout the classification process. Results from the simulations showed that the recommended methods were superior in terms of performance.