Gestures without libraries, toolkits or training: a $1 recognizer for user interface prototypes

Jacob O. Wobbrock, Andrew D. Wilson, and Yang Li. 2007. Gestures without libraries, toolkits or training: a $1 recognizer for user interface prototypes. In Proceedings of the 20th annual ACM symposium on User interface software and technology (UIST '07). ACM, New York, NY, USA, 159-168. DOI=10.1145/1294211.1294238 http://doi.acm.org/10.1145/1294211.1294238

This 2007 paper offers instructions for a simple gesture recognizer that will recognize gestures using a template matching technique. The major goal of the work is to provide prototypers with an easy but somewhat robust way to implement gesture recognition in their applications.

The algorithm is based on resampling, rotating, and translating the stroke data, followed by finding the nearest-neighbor example. The method supports adding new examples to the templates database to allow for different strokes. Online data capture is not necessary, as temporal information is not analyzed. A Golden Section Search is used to find the appropriate rotation.

The authors conducted tests which show the $1 recognizer to be as good as the Rubine algorithm and Dynamic Time Warping. However the algorithm suffers from being unable to detect the differences between circles and ovals or squares and rectangles. Horizontal and vertical lines can also be troublesome, but can be mitigated using a modification to the scaling method.