A new gesture recognition algorithm and segmentation method of Korean scripts for gesture-allowed ink editor

Electronic ink is a form of multimedia data that has recently emerged from the development of pen-based computers which incorporate a stylus pen as the major input device. Gestures have proven to be a useful feature for pen-based user interface. In this paper, we studied some techniques in order to develop a gesture-allowed electronic ink editor for PDAs with hardware constraints. We designed eleven gestures, presented a new feature-based recognition algorithm to identify gestures, and proposed a method for segmenting ink data which mainly consists of Korean scripts into logical units. We implemented GesEdit, the gesture-allowed ink editor on PDAs using the proposed methods. A variety of experiments involving twenty users showed that a gesture recognition rate reaching 99.6% and a correct segmentation rate surpassing 99.8%.