| Summary, etc |
Navarez, Karl Daven., College of Computer Studies, Cebu Institute of Technology-University; March 2012,Touchless Alphanumeric Writer.<br/><br/>Adviser: Prof Jennelyn C. Suson<br/><br/>Existing technology nowadays mostly make use of touch screens. Indeed, these can be very convenient especially when writing with your fingertip. However, the touch screen device itself can be quite expensive to purchase. The device should also be handled with care in order to maintain the sensitivity of the screen. This study presents a spatiaal approach in writing with your fingertip without the burden of holdig device, or wearing a tool on the user's hand such as glove or thimble.<br/><br/>The system uses a single webcam for image capturing. In order to locate the fingertip position, the hand contour is extracted first. A convex hull can be extracted from the contour afterwards. The fingertip location is the farthest edge from the base of the hull. Closed or open fist controls the pen up and down mechanism. This can be achieved using angle detection. During pen-down, the recognizer continuously stores the points of the trajectory of the fingertip and converts it into a series of stokes. Extracted strokes are analyzed by using Dynamic Time Wrap Matching Algorithm and the machine encoded text is outputted after.<br/><br/>Although the system requires some controlled conditions, it can recognize the input strokes. With this, the cost for effective touchless human-computer interaction can be achieved. The study can be improvised in terms of robustness in finger tracking. The usage of a trained classifier is highly recommended since it cannot be affected by illuminaton changes. But it requires thousands of sample images for it to detect effectively. |