A system and process for capturing and implementing handwritten gestures is described. A user may use gestures to quickly command a pen-based input device to perform operations. Using the described embodiments, users may edit text or handwritten ink using a pen in a pen-based computing platform and have the edits affect the text or ink.
1. A method for determining if received strokes are gestures comprising:
receiving at least one stroke as a user input to a computing device;
normalizing said at least one stroke;
computing Chebychev polynomials for coordinates of the at least one stroke;
passing said polynomials to one or more previously computed Bayes nets for known gestures, each of said gestures being associated with a predetermined command to be executed by said computing device;
determining whether said polynomials correlate with one or more of said Bayes nets;
if said polynomials correlate with one or more of said Bayes nets, then causing said computing device to process said at least one stroke as a gesture and execute a corresponding command; and
if said polynomials do not correlate with any of said Bayes nets, then causing said computing device to withhold rendering said at least one stroke as ink until a predetermined plurality of non-correlating strokes have been accumulated, and rendering said plurality of strokes as ink after said plurality of non-correlating strokes have been accumulated.
2. The method according to claim 1 , further comprising:
compressing said at least one stroke into a single point stream; and
passing said single point stream to said one or more Bayes nets.
3. The method according to claim 1 , further comprising:
scaling the time entry of the points of said at least one stroke to create a scaled time; and
passing said scaled time to said one or more Bayes nets.
4. The method according to claim 3 , further comprising:
computing stroke windings and duration of the at least one stroke; and,
passing said scaled time to said one or more Bayes nets.
5. A computer-readable medium having computer-executable instructions for performing steps comprising:
receiving at least one stroke as a user input to a computing device;
normalizing said at least one stroke;
computing Chebychev polynomials for coordinates of the at least one stroke;
passing said polynomials to one or more previously computed Bayes nets for known gestures, each of said gestures being associated with a predetermined command to be executed by said computing device;
determining whether said polynomials correlate with one or more of said Bayes nets;
if said polynomials correlate with one or more of said Bayes nets, then causing said computing device to process said at least one stroke as a gesture and execute a corresponding command; and
if said polynomials do not correlate with any of said Bayes nets, then causing said computing device to withhold rendering said at least one stroke as ink until a predetermined plurality of non-correlating strokes have been accumulated, and rendering said plurality of strokes as ink after said plurality of non-correlating strokes have been accumulated.
6. The computer-readable medium of claim 5 , further comprising computer-executable instructions for performing the steps of:
compressing said at least one stroke into a single point stream; and
passing said single point stream to said one or more Bayes nets.
7. The computer-readable medium of claim 5 , further comprising computer-executable instructions for performing the steps of:
scaling the time entry of the points of said at least one stroke to create a scaled time; and
passing said scaled time to said one or more Bayes nets.
8. The computer-readable medium of claim 7 , further comprising computer-executable instructions for performing the steps of:
computing stroke windings and duration of the at least one stroke; and,
passing said scaled time to said one or more Bayes nets.