EMG pattern classification of normal and anterior cruciate ligament injured subjects

Abstract

The authors have analyzed features of electromyographic (EMG) linear envelopes from six muscles of normal and anterior-cruciate-ligament-injured subjects and found that their phase spectra play major roles in determining the waveform. Pattern classification of six muscles was then done based on such features, and clusterings of some abnormal EMG waveforms were obtained. Several criteria for determining the optimal cluster number were also investigated.<>