Gesture translation system is basically used to find out the meaning of emergency responses from disabled and bed ridden patients by proper interpretation of the gestures shown by them. In order to get pristine gesture components, it is necessary to use a unique gesture without confusion as well as finding the most predominant sensor electrodes which is involved in such gesture activity. An experiment is designed for identifying such predominant sensors, by means of the gestures selected uniquely from activities like lifting a stone, bottle, mobile and book and is tested for signals with five subjects. Analysis was done with the performance of each sensor involved and for each gesture shown by every subject. Results based on the performance of the sensor by highest deviation from the mean are identified as the best contributing channels in classifying the feature. The uniformity in superior performance of those identified electrodes across different gestures and subjects verifies the findings.
