Diagnosis of Neuromuscular Disorders Using Softcomputing Techniques
Akhila Devi B V1, S.Suja Priyadharsini2
1Akhila Devi.B.V, ECE Department, Regional centre of Anna University Tirunelveli Region Tirunelveli, India.
2S.Suja Priyadharsini, ECE Department, Regional centre of Anna University Tirunelveli Region Tirunelveli, India.
Manuscript received on October 23, 2013. | Revised Manuscript received on November 01, 2013. | Manuscript published on November 05, 2013. | PP: 105-111 | Volume-3 Issue-5, November 2013. | Retrieval Number: E1908113513/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Biomedical signals are collection of electrical signals which generated from any organ that signal represents a physical variable of interest. Electromyography (EMG) is a technique for evaluating and recording of electrical activities produced from skeletal muscles. There are so many applications of EMG signals. Major interests lies in the field of clinical as well as biomedical engineering.EMG is used as a diagnostic tool for identifying neuromuscular disorders .Motor unit action potentials (MUPS) provides information about neuromuscular disorders. Traditionally neurophysiologist can access MUPs information from their shapes and patterns using an oscilloscope. But MUPs from different motor neurons will overlap leads to the formation of interference pattern and it is difficult to detect individual shapes accurately. For this reason a number of computer based quantitative EMG analysis algorithm have been developed. In this work, different types of learning methods were used to classify EMG signals. The model automatically classifies EMG signals into normal, myopathy and neuropathy. In order to extract useful information from the EMG signals different feature extraction methods such as discrete wavelet transform(DWT) and auto regressive modeling(AR)are implemented. Adaptive neuro-fuzzy inference system (ANFIS) with hybrid learning algorithm, support vector machine (SVM) and fuzzy support vector machine (FSVM) were compared in relation to their accuracy in the classification of EMG signals. Based on the impacts of features on the EMG signal classification, different results were obtained through analysis of the soft computing techniques.
Keywords: Adaptive neuro-fuzzy inference system (ANFIS), Discrete Wavelet Transform (DWT), Electromyography (EMG) Fuzzy SVM (FSVM), Support vector machine (SVM).