STATISTICAL FEATURES BASED DECISION TREE APPROACH FOR HAND GESTURE RECOGNITION

Abstract

The increasing trend of simulating the human intelligence into machine intelligence has resulted into an emphasis on the advancements in computer vision; one of the major areas in computer vision is gesture recognition. Gesture recognition is a very important task which can be used in multiple applications and system automation. Especially in virtual reality applications gesture recognition is of significant importance. Approaching towards the social importance of gesture recognition it can rightly be said that gesture recognition has solved a lot many social problems, for example the vocally and hearing impaired people get socially isolated due to the communicational gap in between the normal people and them. The deaf and dumb people need not only learn the standard sign language but the core issue is that they can communicate with the normal people of society. It is also not possible for all the normal people that they learn the sign language to understand whatever is said through gestures. So the communicational gap still stays there even after teaching deaf and dumb people with sign language.There is need of development of such robust systems that facilitate communication between them and provide them with easy to operate communication system. This research presents a study of multiple gesture recognition techniques and implementation of a new proposed English alphabet gesture recognition system which is based on implementation of decision trees that are fed with the features of the gesture to be recognized. First of all, the hand segmentation is discussed; the technique used for hand segmentation in this research is skin color based technique in which the HSV and YCbCr color spaces have been used. Then the statistical features of the detected hand showing the gesture has been extracted and fed to the decision tree. The decision tree based rules result in the classification of the input gesture as the relevant alphabet