Windows 8 Touch-liked Gesture Using Computer Vision Technique on Webcam

Abstract

Number of tablet users is rapidly increasing due to the technologies. Windows gesture is required to allow normal laptop users (without touch screen) touch or tap within Windows 8 modern UI applications. However, a gesture detected device, which is included as a part of common laptops, i.e. Web camera is the challenge of providing tablet-liked user experience. This paper shows how to improve normal web camera for touch-like gestures with these following. First, non-human skin pixels are subtracted from each Webcam image. Then, determining hand from facial removed image by applying Haar-like feature. Separating each gesture from compactness and protrusion, and blinding them with suitable Windows command. Therefore, we achieve Windows touch-liked gesture program. Comparing result with other programs shows that our method is precise but over-consuming resources. Our method provides alternative idea of combined computer vision techniques for Windows touch-like gestures.