PCA based shape recognition for capacitive touch display

With the growing diffusion of touch screen based consumer devices, the development of algorithms able to discriminate among different shapes obtained by touching the display becomes very important. For instance, the detection and recognition of fingers represent fundamental information in many touch based user applications such as image zooming, swipe, figure rotation, gaming, finger painting, etc. These algorithms are also extremely useful to recognize accidental touches in order to avoid involuntary activation of the device functionalities. Moreover, the recognition engine should be able to identify shapes acquired in critical conditions (e.g. slightly wet fingers). In this paper we present both hardware and software components to recognize different classes of shapes: single finger, double fingers and palm.