Originally this was done using a continuous observation node,
but I decided that was too complicated for a first version of the
system, so I discretized the observation space.
Originally this was done using a continuous observation node,
but I decided that was too complicated for a first version of the
system, so I discretized the observation space.
Make a new stroke set that is rescaled so that the aspect ratio of the
bounding box, of arbitrary orientation, remains the same and the longest
axis of the new bounding box has the specified length.
Method used to return the covariance matrix for the stroke list -
NOTE: computeCovarianceMatrix() should be called before getting the covariance matrix
The fudge factor determines how close to the border between
two speech events a gesture event has to be in order to be
counter as part of the feature vector.
Get an initial set of match clusters based on points matching under the
costThreshold, that satisfy the distance constraint defined by the
multipler and within the given limits for orientation hypothesis.
Make sure that you synchronize on the index returned by this
method if you are doing multiple modifications to it so that
someone doesn't try to read from it while you are changing it.
Returns the maximum angle (in polar coordinate space) of this symbol
If the symbol crosses the 0-2Pi boundary, this may not return the "maximum" value you expect