Electromyogram based automatedHand gesture recognition systemUsing signal processing techniquesAnd artificial neural networks

Abstract

The biomedical signals are described as the collection of electricalsignals acquired from any organ and this signal is normally a function of time.These electrical biomedical signals refer to the change in electric currentproduced by the sum of an electrical potential difference across a particulartissue or organ. These signals are of very low amplitude and low frequencyelectrical signals. The biomedical signals have been divided into groupsaccording to their inherent characteristics. Among the biomedical signals, thesignal used for hand gesture recognition is Electromyography (EMG) signal.It is the recording of the electrical potential retrieved from the muscle fibers.The surface electrodes are employed to pick up the EMG signals from thebody. When the EMG signal is retrieved from the muscle fiber, noises addwith them as they travel along different tissues.With the immense growth of communication technology, humantend to use hand gestures in their communication process to explain theirinnovations and motivations. Thus, hand gesture recognition is considered tobe an important part of Human Computer Interaction (HCI). It givescomputers the capacity of capturing and interpreting hand gestures andexecuting commands accordingly.Capturing human hand gesture by the computer is termed as hand newlinegesture acquisition. This can be done using vision-based hand gesturerecognition and data glove-based technique. The vision-based techniques arestatic hand and real-time hand gesture recognition techniques. The data glovetechnique requires five flex sensors to be attached to the glove. The visionbasedtechnique is more stable and reliable compared to the data glove-basedtechnique. But it is very difficult to design a vision based interface for generalusage. Again newline newline