Partial overlap and delayed stroke input recognition
View Patent ↗An optimal recognition for handwritten input based on receiving a touch input from a user may be selected by applying both a delayed stroke recognizer as well as an overlapping recognizer to the handwritten input. A score may be generated for both the delayed stroke recognition as well as the overlapping recognition and the recognition corresponding to the highest score may be presented as the overall recognition.
1. A computer-implemented method comprising:
receiving a sequence of strokes that represent a handwritten input;
providing the sequence of strokes to both a delayed stroke handwriting recognizer and an overlapping handwriting recognizer,
wherein a delayed stroke handwriting recognizer is a handwriting recognizer that orders strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes,
wherein an overlapping handwriting recognizer is a handwriting recognizer that does not reorder strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes;
receiving, from each of the delayed stroke handwriting recognizer and the overlapping handwriting recognizer, (i) a set of one or more candidate letters that the handwriting recognizer has selected as corresponding to the sequence of strokes, and (ii) a confidence score associated with the selection; and
selecting, from among the set of candidate letters from the delayed stroke handwriting recognizer and the set of candidate letters from the overlapping handwriting recognizer, a particular set of candidate letters based on the confidence scores.
2. The method of claim 1 , wherein the sequence of strokes is received through a touch sensitive surface.
3. The method of claim 1 , wherein one or more candidate letters from the set of candidate letters overlap each other.
4. The method of claim 1 , wherein the sequence of strokes are received in an order opposite a conventional reading direction.
5. The method of claim 1 , wherein the delayed stroke handwriting recognizer analyzes a portion sequence of strokes that is received over a particular period of time.
6. The method of claim 5 , wherein the particular period of time is based on a predetermined amount of time, user-provided training handwriting input, or previously received sequence of stokes.
7. The method of claim 1 , wherein the overlapping handwriting recognizer generates the confidence score based on a portion sequence of strokes that is based on a period of time, positional information, and previously received sequence of strokes.
8. The method of claim 1 , wherein the overlapping handwriting recognizer and the delayed stroke handwriting recognizer exchange data related to feature extraction and partial recognition of a portion of letter.
9. The method of claim 1 , comprising:
generating a graph that includes nodes that correspond the sets of one or more candidate letters from each handwriting recognizer,
wherein selecting the particular set of candidate letters is further based on a least costly path that traverses the graph.
10. The method of claim 1 , wherein the confidence score for each of the overlapping handwriting recognizer and the delayed stroke handwriting recognizer is based at least in part on a similarity between a portion of the sequence of strokes and handwriting characters and on confidence scores of a previously received portion of the sequence of strokes.
11. A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
receiving a sequence of strokes that represent a handwritten input;
providing the sequence of strokes to both a delayed stroke handwriting recognizer and an overlapping handwriting recognizer,
wherein a delayed stroke handwriting recognizer is a handwriting recognizer that orders strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes,
wherein an overlapping handwriting recognizer is a handwriting recognizer that does not reorder strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes;
receiving, from each of the delayed stroke handwriting recognizer and the overlapping handwriting recognizer, (i) a set of one or more candidate letters that the handwriting recognizer has selected as corresponding to the sequence of strokes, and (ii) a confidence score associated with the selection; and
selecting, from among the set of candidate letters from the delayed stroke handwriting recognizer and the set of candidate letters from the overlapping handwriting recognizer, a particular set of candidate letters based on the confidence scores.
12. The system of claim 11 , wherein the sequence of strokes is received through a touch sensitive surface.
13. The system of claim 11 , wherein one or more candidate letters from the set of candidate letters overlap each other.
14. The system of claim 11 , wherein the sequence of strokes are received in an order opposite a conventional reading direction.
15. The system of claim 11 , wherein the delayed stroke handwriting recognizer analyzes a portion sequence of strokes that is received over a particular period of time.
16. The system of claim 15 , wherein the particular period of time is based on a predetermined amount of time, user-provided training handwriting input, or previously received sequence of stokes.
17. The system of claim 11 , wherein the overlapping handwriting recognizer generates the confidence score based on a portion sequence of strokes that is based on a period of time, positional information, and previously received sequence of strokes.
18. The system of claim 11 , wherein the overlapping handwriting recognizer and the delayed stroke handwriting recognizer exchange data related to feature extraction and partial recognition of a portion of letter.
19. The system of claim 11 , wherein the operations further comprise:
generating a graph that includes nodes that correspond the sets of one or more candidate letters from each handwriting recognizer,
wherein selecting the particular set of candidate letters is further based on a least costly path that traverses the graph.
20. The system of claim 11 , wherein the confidence score for each of the overlapping handwriting recognizer and the delayed stroke handwriting recognizer is based at least in part on a similarity between a portion of the sequence of strokes and handwriting characters and on confidence scores of a previously received portion of the sequence of strokes.
21. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
receiving a sequence of strokes that represent a handwritten input;
providing the sequence of strokes to both a delayed stroke handwriting recognizer and an overlapping handwriting recognizer,
wherein a delayed stroke handwriting recognizer is a handwriting recognizer that orders strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes,
wherein an overlapping handwriting recognizer is a handwriting recognizer that does not reorder strokes according to their respective horizontal spatial location before selecting one or more characters that correspond to the strokes;
receiving, from each of the delayed stroke handwriting recognizer and the overlapping handwriting recognizer, (i) a set of one or more candidate letters that the handwriting recognizer has selected as corresponding to the sequence of strokes, and (ii) a confidence score associated with the selection; and
selecting, from among the set of candidate letters from the delayed stroke handwriting recognizer and the set of candidate letters from the overlapping handwriting recognizer, a particular set of candidate letters based on the confidence scores.
22. The medium of claim 21 , wherein the sequence of strokes is received through a touch sensitive surface.
23. The medium of claim 21 , wherein one or more candidate letters from the set of candidate letters overlap each other.
24. The medium of claim 21 , wherein the sequence of strokes are received in an order opposite a conventional reading direction.
25. The medium of claim 21 , wherein the overlapping handwriting recognizer generates the confidence score based on a portion sequence of strokes that is based on a period of time, positional information, and previously received sequence of strokes.
26. The medium of claim 21 , wherein the operations further comprise:
generating a graph that includes nodes that correspond the sets of one or more candidate letters from each handwriting recognizer,
wherein selecting the particular set of candidate letters is further based on a least costly path that traverses the graph.
27. The medium of claim 21 , wherein the confidence score for each of the overlapping handwriting recognizer and the delayed stroke handwriting recognizer is based at least in part on a similarity between a portion of the sequence of strokes and handwriting characters and on confidence scores of a previously received portion of the sequence of strokes.