A hand gesture recognition technique from real time video

Hand gesture recognition plays a vital role in human computer interaction. It is a difficult task to classify a real time video. This paper presents a dynamic hand gesture recognition method from a real time video of hand gesture. The colour space `YC_bC_r´ is used for correctly determining the skin colour. Features from real time video are then extracted using Histogram of Oriented Gradient (HOG) descriptors. The features are then entered in a support vector machine (SVM) classifier for classification purpose. The gesture recognition has been done for bare hand.