| Summary: | 碩士 === 實踐大學 === 資訊科技與管理學系碩士班 === 100 === Touch screen has been widely used as user interface for most mobile embedded systems. Currently, both the well-known iOS and Android operating systems have been fully supporting multi-touch capabilities. But most mobile applications run on them are still based on only single or two strokes of touch gestures to make a simple tap, draw, or zoom. In this paper, we propose a new gesture recognizer to track the related attributes of the positions, gap time, sequence, directions, and styles (such as dot, line, arc, or circle) among touch strokes to realize a richer gestures and customizable design of recognition. For example, we can customize the gesture of drawing a cross X with two strokes corresponding to do the action of closing the running application immediately. The recognizer designed with a two-level finite-state machine can be embedded to any applications of iOS to fully recognize all possible four kinds of gestures: single-touch and single-stroke, single-touch and multi-stroke, multi-touch and single-stroke, and
multi-touch and multi-stroke. We also support users to customize their favorite gestures according to operations of applications and need not to rebuild them. Since most hand-writing recognition algorithms are CPU-bound, our approach will keep track the related attributes among strokes to recognize gesture within a customized state transition diagram such that our proposed model not only a simple and efficient power-saving algorithm design, but also a personal customized gesture recognizer.
|
|---|
