Smart Human Computer Interface with EMG and Vision Based on Multi-modal Information Fusion

A smart Human-Computer Interface(HCI)replacing conventional mouse interface is proposed.The interface is able to control cursor and command action with only hand.Four finger motions(left click,right click,hold,drag)are used to command the interface.Also the authors materialize cursor movement control using image processing.The measure what they use for inference is entropy of Electromyogram(EMG)signal,Gaussian modeling and maximum likelihood estimation.In image processing for cursor control,they use color recognition to get the center point of finger tip from marker,and map the point onto cursor.Accuracy of finger movement inference is over 95% and cursor control works naturally without delay.They materialize whole system to check its performance and utility.