Last week at the ICMI 2012 conference, I presented a new gesture recognizer in the $-family, called $P. $P is highly accurate, needing few training examples or templates, and is able to handle gestures made with any number of strokes in any order or direction, but uses simple concepts that make it accessible to those other than experts in machine learning or pattern matching. Find my presentation slides here.
Tag Archives: icmi 2012
Co-authors Radu-Daniel Vatavu and Jacob O. Wobbrock and I have had a paper accepted to ICMI 2012, titled “Gestures as Point Clouds: A $P Recognizer for User Interface Prototypes,” in which we introduce $P, the latest member of the $-family of gesture recognizers. $P can handle multistroke and unistroke gestures alike with high accuracy, and remedies the main limitations of $N in terms of cost to store and match against all possible multistroke permutations.
Here is the abstract:
Rapid prototyping of gesture interaction for emerging touch platforms requires that developers have access to fast, simple, and accurate gesture recognition approaches. The $-family of recognizers ($1, $N) addresses this need, but the current most advanced of these, $N-Protractor, has significant memory and execution costs due to its combinatoric gesture representation approach. We present $P, a new member of the $-family, that remedies this limitation by considering gestures as clouds of points. $P performs similarly to $1 on unistrokes and is superior to $N on multistrokes. Specifically, $P delivers >99% accuracy in user-dependent testing with 5+ training samples per gesture type and stays above 99% for user-independent tests when using data from 10 participants. We provide a pseudocode listing of $P to assist developers in porting it to their specific platform and a “cheat sheet” to aid developers in selecting the best member of the $-family for their specific application needs.