APPARATUS AND METHOD FOR GESTURE RECOGNITION USING MULTICLASS SUPPORT VECTOR MACHINE AND TREE CLASSIFICATION

摘要

Disclosed is a device for recognizing a gesture using a multi-class support vector machine (SVM) and tree classification. A three-dimensional information extracting unit extracts three-dimensional information on a user joint based on a gesture of a user which is inputted from a kinect sensor. A gesture classification tree designing unit designs a gesture classification tree in advance according to a basic pattern of a gesture to classify the extracted three-dimensional information according to the gesture classification tree. A chain code feature extracting unit extracts a feature vector of a chain code for the gesture. A gesture classifying unit classifies the gesture after learning by the multi-class SVM by using the chain code and histogram information for the gesture as the feature vector. According to the present invention, the device is able to accurately recognize a gesture through a natural user interface.