Specifying Gestures by Example

Dean Rubine. 1991. Specifying gestures by example. In Proceedings of the 18th annual conference on Computer graphics and interactive techniques (SIGGRAPH '91). ACM, New York, NY, USA, 329-337. DOI=10.1145/122718.122753 http://doi.acm.org/10.1145/122718.122753

This 1991 paper offers a simple feature based method for detecting single-stroke gestures. The paper provides 13 features, two which are optional (based on temporal information), and a simple linear classifier. The features are derived from single stroke input (with temporal information) using simple geometric reasoning techniques. The features were derived from empirical observation of what worked well. The GRANDMA toolkit (an implementation) allows users to teach new gestures to the classifier.

The paper admits that more work must be done to distinguish certain gestures, for which the 13 features are not enough. Additionally, the methods described may be extended to support eager classification or multi-finger recognition.

Overall a very interesting paper for its time. Defines a simple and easy to use method for basic gesture recognition which uses temporal information.