Human Computer Interface for Gesture-Based Editing System

The use of hand gesture provides an attractive alternative to cumbersome interface devices for human-computer interactio n(HCI). Many methods for hand gesture recognition using visual analysis have been proposed such as syntactical analysis, neural network(NN), Hidden Markov Model(HMM) and so on. In our research, a HMM is proposed for alphabetical hand gesture recognition. In the preprocessing stage, the proposed approach consists of three different procedures for hand localization, hand tracking and gesture spotting. The hand location procedure detects the candidated regions on the basis of skin-color and motion in an image.. The hand tracking algorithm finds the centroid of a moving hand region, connect those centroids, and thus, produces a trajectory. The spotting algorithm divides the trajectory into real and meaningless gestures. In constructing a feature database, the proposed approach use the location, angle and velocity feature code, and employ a k-means algorithm for codebook of HMM. In our experiments, 2400 trained gestures and 2400 untrained gestures are used for training and testing, respectively. Those experimental results demonstrate that the proposed approach yields a higher and satisfying recognition rate with various gestures. Hand gesture recognition using visual devices has a number of potential application in HCI (human computer interaction), VR(virtual reality), machine control in the industry field, and so on[1,2]. Most conventional approaches to hand gesture recognition has employed external devices such as datagloves, maker and so on. But, for more natural interface, hand gesture must be recognized from visual images without any external devices. Many methods for hand gesture recognition using visual device have been proposed such as syntactical analysis, neural based approach, HMM (hidden markov model) based recognition[3,4]. As gesture is the continuous motion on the sequential time series, HMM must be a prominent recognition tool. Several hand gesture recognition systems have been developed using various features computed from static images or image sequences[5]. Segan[6] used edge-based technique to extract image parameters from simple silhouettes and developed a system which can recognize 10 distinct pose in real-time. Hunter[7] used the Zernike moments as the image features and developed a system in which the sequence of hand gesture were recognized using HMM. Starner[3] used image geometry parameter as the image features and employed a HMM five states topology for the gesture classification. In our research, we consider the planar hand gesture in front of camera and detect 16-dimension location codes as input vectors for the HMM network. We use …