Gesture recognition can be difficult, arcane, and specialized.
But it doesn't have to be. Published in 2007, the $1 recognizer, and
the $-family recognizers that followed, made stroke-gesture
recognition easy to add to any user interface.
The $-family recognizers, of which
$1,
$N,
$P, and
$Q
are canonical members, along with enhancements
Protractor and
$P+, have
had significant impact on the use of gestures in
user interface prototypes and
interactive systems.
The $1 paper and its siblings
led to numerous follow-ons by other researchers,
leading to the "extended $-family" of
technologies. All adopt the same philosophy espoused
in the original $1 paper: to take what
are typically complex algorithms
best understood by specialists and make them easy to
convey, implement, and deploy on any platform for
non-specialists whose objective is to quickly enhance
interactivity. Accordingly, the $1 paper began a trend,
followed by other $-family publications, of putting their
entire pseudocode necessary for implementation directly in
their papers.
Some of the core $-family publications were recognized with awards, including:
Microsoft used $N to recognize multistroke gestures on their
PocketTouch
device, which enabled gesture-sensing through a pants pocket for quick, ambient interactions.
German firm Prefrontal Cortex employed
$P for gesture control in
Ombre Fabula,
a gesture-controlled interactive shadow play. The prototype won 2nd place at the
2014 Intel RealSense App Challenge.
$1's direct successor, Protractor, sped up $1's iterative angular
alignment process using a closed-form calculation based on inverse cosine distance, and became the basis for Google's
android.gesture package.
$1 and the other canonical $-family recognizers inspired other researchers to develop follow-ons that share
the $-family motivation. In chronological order:
Over the years, many developers have sent in their own implementations of our $-family recognizers in various programming languages.
We make no representations as to the correctness or completeness of these implementations, but offer them here "as is." (We also don't
check these links very often, so some may have gone stale.)
(Have one to add? Email
wobbrock@uw.edu.)