Touchware: a software-based technique for high-resolution multi-touch sensing devices

Abstract

Finger pointing on touch screens is a very natural way of human computer interaction. However, for many capacitive touch sensing devices, it may suffer from the nature of direct input since the size of human fingers and the lack of sensing precision make absolute positioning on touch screen difficult, especially on multi–touch sensing devices. Even if high–resolution/precision multi–touch devices become popular in the market, the cost is high and the positioning algorithm is not a flexible module. In this paper, we present Touchware, a software–based technique to overcome these limitations with low cost and to provide support for the development of multi–touch applications for rich input modalities. We introduce the maxima–based clustering algorithm and weight–based geometric algorithm for accurate finger positioning and first–contact–based decision with sniper for ghost pattern elimination. We evaluated the performance and show how the techniques can be successfully used for single–touch and multi–touch applications.