Sub-vocal phoneme-based EMG pattern recognition and its application in diagnosis

Silent Communication Device

“…Similar devices have been made as shown in projects such as [23] and [33] in which said projects employ wavelet analysis (explained in much greater detail later in chapter 3) in order to effectively process subvocal speech to allow said subvocal speech to be more easily classified as its intended speech for the final output. The wavelet analysis allows these devices to classify subvocal speech up to accuracies between 70% and 80%, but only through the use of expensive and not pocket-sized equipment.…”

“…Khan [33] which feature the extraction of EMG subvocal signals from the throat, amplification and filtering of said signals to allow them to be large enough in amplitude to be accurately sampled without noise, wavelet analysis to extract features from the EMG signals, and classification using a neural network in order to classify the signals to accuracies between 70% to 80%, showing a relatively simple way of translating subvocal speech into a range of words, vowels, or phonemes with decent performance. In addition, the paper "MSP430 Implementation of Wavelet Transform for Purposes of Physiological Signals Processing," by R. Stojanović and S. Knežević [31] demonstrates an example of a microcontroller being used to employ wavelet analysis on signals extracted from the human body, showing that wavelet analysis through use of a microcontroller is possible, while the paper "Subvocal Speech Recognition Based on EMG Signal Using Independent Component Analysis and Neural Network MLP," by J.A.G.…”

“…Although other means of extracting features from subvocal signals exist such as using Independent Component Analysis instead of wavelet analysis [3] that give much more accurate classification results, these methods are much more complex than that of wavelet analysis and thus would require much more complex programs to implement them, which microcontrollers are unlikely to be able to process without at least very slow or even poor performance, while [31] has proven that wavelet analysis works with microcontrollers effectively. Combining the wavelet analysis through use of a microcontroller [31] (using the more advanced microcontroller MSP432P401R instead of a MSP430 microcontroller), the amplification and filtering circuit of [3], and the subvocal speech translation methods of [23] and [33], the Silent Communication Device could be brought to fruition.…”

“…In order to capture and classify these small vibrations as usable data, the small, collective electric signals controlled by the nervous system that are produced during muscle contraction (EMG (Electromyograpy) signals) can be captured as they pass through the muscles that vibrate the vocal cords during subvocalization [25] via EMG electrodes attached to these muscles. The signal bandwidth in which significant EMG activity occurs that is the most useful in terms of classification and usable data is from 5 Hz to 450 Hz [26], while amplitudes range on average from about 0 to 600, occasionally spiking up to about 1000 µV with Ag-AgCl surface EMG electrodes [33].…”

“…The first and most significant approach in terms of influence on the design of the neural network was the paper "Sub-vocal Phoneme-Based EMG Pattern Recognition and its application in Diagnosis" by Mosarrat Jahan and Munna Khan [33]. In this paper, EMG signals were obtained using EMG electrodes on the throat sampled at 500 samples/ second.…”