Granted Patent B2

US 12,282,737 · App. 17/851,275 · Granted Apr 22, 2025

Inventors: Steven N. Bathiche (Bellevue, WA); Moshe R. Lutz (Bellevue, WA)

Assignee: Microsoft Technology Licensing, LLC

G06F40/274G06F40/171G06F40/30G06V30/19147G06V30/347

Abstract

Claims (51)

1. A system for generating predicted ink strokes comprising:

at least one processor;

memory storing instructions that, when executed by the at least one processor, cause the system to perform a set of operations, the set of operations comprising:

receiving ink stroke data, wherein the received ink stroke data comprises a plurality of ink stroke data points and writing pressure associated with the ink stroke data points, and wherein the ink stroke data points form information corresponding to penmanship cleanliness and hand tilt of a user;

inputting the ink stroke data and a semantic context into a model, wherein the model is trained based on ink writing samples to predict ink strokes including a plurality of ink stroke data points that replicate a writing pressure, penmanship cleanliness, and hand tilt of a user;

determining, from the trained model, one or more predicted ink strokes, wherein the determining comprises:

generating a plurality of predicted ink strokes, wherein the plurality of predicted ink strokes each comprise a plurality of predicted ink stroke data points that replicate the writing pressure, penmanship cleanliness, and hand tilt of the user;

ranking the plurality of predicted ink strokes based on the semantic context;

selecting the one or more predicted ink strokes from the plurality of predicted ink strokes, based on the ranking; and

generating an indication of the one or more predicted ink strokes, wherein the indication of the one or more predicted ink strokes includes the replicated writing pressure, penmanship cleanliness, and hand tilt of the user; and

causing the one or more predicted ink strokes that include the replicated writing pressure, penmanship cleanliness, and hand tilt of the user to be displayed.

2. The system of claim 1 , wherein the model is a trained machine-learning model.

3. The system of claim 2 , wherein the ink stroke data and the semantic context are automatically input into the trained machine-learning model, as the ink stroke data is received.

4. The system of claim 2 , wherein the trained machine-learning model is a neural network that is trained based on the ink writing samples, the ink writing samples being received from a data set.

5. The system of claim 1 , wherein the received ink stroke data comprises a partial stroke input.

6. The system of claim 1 , wherein the received ink stroke data comprises a full stroke input, the full stroke input corresponding to an alphanumeric character.

7. The system of claim 1 , wherein the received ink stroke data comprises one or more of letter spacing, sentence spacing, and writing orientation.

8. A system for generating predicted ink strokes comprising:

at least one processor;

memory storing instructions that, when executed by the at least one processor, cause the system to perform a set of operations, the set of operations comprising:

receiving ink stroke data, wherein the ink stroke data comprise a plurality of ink stroke data points and writing pressure associated with the data point, and wherein the ink stroke data points form information corresponding to penmanship cleanliness and hand tilt of a user:

inputting the ink stroke data and a semantic context into a trained machine-learning model, wherein the trained machine-learning model is trained based on ink writing samples to predict ink strokes including a plurality of ink stroke data points that replicate a writing pressure, penmanship cleanliness, and hand tilt of a user;

determining, from the trained machine-learning model, two or more predicted ink strokes, wherein the determining comprises:

generating a plurality of predicted ink strokes, wherein the plurality of predicted ink strokes each comprise a plurality of predicted ink stroke data points that replicate the writing pressure, penmanship cleanliness, and hand tilt of the user;

ranking the plurality of predicted ink strokes based on semantic context; and

selecting the two or more predicted ink strokes from the plurality of predicted ink strokes, based on the ranking; and

generating a plurality of indications that each correspond to a different predicted ink stroke from the two or more predicted ink strokes, wherein the plurality of indications include the replicated writing pressure, penmanship cleanliness, and hand tilt of the user;

causing the two or more predicted ink strokes that include the replicated writing pressure, penmanship cleanliness, and hand tilt of the user to be displayed;

receiving a selection of one of the plurality of indications; and

updating the trained model based on the selection, the input ink stroke data, and the input semantic context, to accommodate the user.

9. The system of claim 8 , wherein the set of operations further comprise:

displaying the one of the predicted ink strokes that corresponds to the selected one of the plurality of indications.

10. The system of claim 8 , wherein the ink stroke data and the semantic context are automatically input into the trained machine-learning model, as the ink stroke data is received.

11. The system of claim 8 , wherein the trained machine-learning model is a neural network that is trained based on the ink writing samples, the ink writing samples being received from a data set.

12. The system of claim 8 , wherein the received ink stroke data comprises a partial stroke input.

13. The system of claim 8 , wherein the received ink stroke data comprises a full stroke input, the full stroke input corresponding to an alphanumeric character.

14. The system of claim 8 , wherein the ink stroke data comprises one or more of letter spacing, sentence spacing, and writing orientation.

15. A method for generating predicted ink strokes, the method comprising:

receiving ink stroke data, wherein the received ink stroke data comprises a plurality of ink stroke data points and writing pressure associated with the ink stroke data points, and wherein the ink stroke data points form information corresponding to penmanship cleanliness and hand tilt of a user;

inputting the ink stroke data and a semantic context into a trained machine-learning model, wherein the trained machine-learning model is trained based on ink writing samples to predict ink strokes including a plurality of ink stroke data points that replicate a writing pressure, penmanship cleanliness, and hand tilt of a user;

determining, from the trained machine learning model, a predicted ink stroke, wherein the determining comprises:

generating a plurality of predicted ink strokes, wherein the plurality of predicted ink strokes each comprise a plurality of predicted ink stroke data points that replicate the writing pressure, penmanship cleanliness, and hand tilt of the user;

sorting the plurality of predicted ink strokes based on the semantic context; and

selecting the predicted ink stroke from the plurality of predicted ink strokes, based on the sorting;

generating an indication of the predicted ink stroke, wherein the indication of the predicted ink stroke includes the replicated writing pressure, penmanship cleanliness, and hand tilt of the user; and

causing the predicted ink stroke that includes the replicated writing pressure, penmanship cleanliness, and hand tilt of the user to be displayed.

16. The method of claim 15 , wherein the ink stroke data and the semantic context are automatically input into the trained machine-learning model, as the ink stroke data is received.

17. The method of claim 15 , wherein the received ink stroke data comprises a partial stroke input.

18. The method of claim 15 , wherein the predicted ink stroke comprises two or more predicted ink strokes selected from the plurality of predicted ink strokes, wherein the indication comprises a plurality of indications each corresponding to a respective predicted ink stroke of the two or more predicted ink strokes, and wherein the method further comprises:

receiving a selection of one of the plurality of indications; and

updating the trained model based on the selection, the input ink stroke data, and the input semantic context, to accommodate the user.

Assignments (1)

Continuity (1)

References Cited (16)