Image analysis – Image segmentation – Segmenting individual characters or words
Reexamination Certificate
2000-08-22
2004-06-22
Chang, Jon (Department: 2623)
Image analysis
Image segmentation
Segmenting individual characters or words
C382S186000, C382S187000, C382S218000, C382S229000
Reexamination Certificate
active
06754386
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to computer devices, and more particularly computer devices arranged to receive handwritten input.
BACKGROUND OF THE INVENTION
Contemporary computing devices allow users to enter handwritten words (e.g., in cursive handwriting and/or printed handwritten characters) and symbols (e.g., a character in Far East languages). The words and symbols can be used as is, e.g., to function as readable notes and so forth, or can be converted to text for more conventional computer uses. To convert to text, for example, as a user writes strokes representing words or other symbols onto a touch-sensitive computer screen or the like, a handwriting recognizer (e.g., trained with millions of samples, employing a dictionary, context and other rules) is able to convert the handwriting data into dictionary words or symbols. In this manner, users are able to enter textual data without necessarily needing a keyboard.
However, contemporary recognizers are too processor-intensive to use continuously, particularly with small hand-held (pocket-sized) personal computers. Instead, the devices are configured such that users can request recognition when desired, or such that the recognizer operates only in the background so that the users do not have to wait for recognition on a word-by-word basis. Moreover, some recognizers are configured to use the context of surrounding words or symbols to help determine other words or symbols, and thus waiting to recognize groups of words often provides more accurate recognition.
At the same time, however, the user may wish to edit words and symbols before they have been recognized, e.g., delete a word, insert a word between two others, format a word, move words around, and so forth, such as by cutting and pasting. Also, the system needs to know when to wrap handwritten data to the next line, and needs to maintain this information as the user zooms the display in and out. As such, some relatively lightweight, initial ink processing is performed to parse the input data into various segments, without involving the recognizer. This initial processing may be orders of magnitude faster than full recognition, but has problems, however, in that because it only looks for limited characteristics via some fuzzy rules, the distinct segments initially determined may not actually correspond to the distinct words or symbols that the user intended. Thus, as a user edits words and zooms in and out, the user may undesirably find that what was intended to be one word is treated as two, two words treated as one, and so forth.
SUMMARY OF THE INVENTION
Briefly, the present invention provides a method and system that corrects for parser segmentation errors by sending an entire line of ink to a recognizer, and then comparing, on a word-by-word basis, the initial segmentation guesses of the parser with the more-thoroughly recognized segmentation results of the handwriting recognition engine. In the correction process, the ink words are efficiently adjusted with relatively little data manipulation. More particularly, the recognizer is fed a series of strokes on a line, in order, from which the recognizer returns its segmentation information. For ink word breaks that are the same between the ink parsing and the recognizer for any given set of data, the existing ink word is unchanged. For ink words that are recognized differently, a new ink word is created and the handwriting (including stroke) data of the parser's ink word is manipulated to create a new ink processor word (or words) to match the recognizer output.
The present invention thus retains the desirability of rapidly grouping sets of strokes into segmented ink word units as they are written, without first requiring complete recognition, for example for word processing (editing and zooming) purposes, and so forth. However, once recognition is performed, the segmentation performed by the lightweight parser, which is generally not as accurate as the segmentation performed by the recognizer, is efficiently and flexibly corrected by a data manipulation process to match the recognizer's output.
Other advantages will become apparent from the following detailed description when taken in conjunction with the drawings, in which:
REFERENCES:
patent: 5594640 (1997-01-01), Capps et al.
patent: 5615285 (1997-03-01), Beernink
patent: 5675665 (1997-10-01), Lyon
patent: 5682439 (1997-10-01), Beernink et al.
patent: 5710916 (1998-01-01), Barbara et al.
patent: 5764799 (1998-06-01), Hong et al.
patent: 5812697 (1998-09-01), Sakai et al.
patent: 5828991 (1998-10-01), Skiena et al.
patent: 5855000 (1998-12-01), Waibel et al.
patent: 5857099 (1999-01-01), Mitchell et al.
patent: 5953451 (1999-09-01), Syeda-Mahamood
patent: 5960447 (1999-09-01), Holt et al.
patent: 5995664 (1999-11-01), Shimomura
patent: 6005973 (1999-12-01), Seybold et al.
patent: 6014460 (2000-01-01), Fukushima et al.
patent: 6041141 (2000-03-01), Yamamoto et al.
patent: 6111985 (2000-08-01), Hullender et al.
patent: 6473517 (2002-10-01), Tyan et al.
patent: 6493464 (2002-12-01), Hawkins et al.
patent: 2002/0013795 (2002-01-01), Dresevie et al.
patent: 2002/0049796 (2002-04-01), Dresevie et al.
patent: 07110845 (1995-04-01), None
Liang et al. “Efficient Algorithms for Segmentation and Recognition of Printed Characters in Document Processing.” IEEE Pacific Rim Conf. on Communications, Computers and Signal Processing, vol. 1, May 19, 1993, pp. 240-243.*
Lee et al. “A New Methodology for Gray-Scale Character Segmentation and Recognition.” Proc. of the 3rdInt. Conf. on Document Analysis and Recognition, vol. 1, Aug. 14, 1995, pp. 524-527.*
Ariki et al. “Segmentation and Recognition of Handwritten Characters Using Subspace Method.” Proc. of the 3rd Int. Conf. on Document Analysis and Recognition, vol. 1, Aug. 14, 1995, pp. 120-123.*
Zhao et al. “A New Method for Segmenting Unconstrained Handwritten Numeral String.” Proc. of the 4thInt. Conf. on Document Analysis and Recognition, vol. 2, Aug. 18, 1997, pp. 524-527.*
Arica et al. “A New Scheme for Off-Line Handwritten Connected Digit Recognition.” Proc. KES '98, 2ndInt. Conf. on Knowledge-Based Intelligent Electronic Systems, vol. 2, Apr. 21, 1998, pp. 329-335.*
Jung et al. “Machine Printed Character Segmentation Method Using Side Profiles.” Proc. SMC '99, IEEE Int. Conf. on Systems, Man, and Cybernetics, vol. 6, Oct. 12, 1999, pp. 863-867.*
Senda et al. “Document Image Retrieval System Using Character Candidates Generated by Character Recognition Process.” Proc. Of the 2ndInt. Conf. on Document Analysis and Recognition, Oct. 20, 1993, pp. 541-546.
Carrano. Data Abstraction and Problem Solving with C++; Walls and Mirrors. The Benjamin/Cummings Publishing Company, Inc,. 1995, pp. 150-163, 175-176, 179-180.
Morita et al. “A Fast and Compact Data Structure of Storing Multi-Attribute Relations Among Words.” IEEE In Conf. on Systems, Man and Cybernetics, vol. 3, Oct. 11, 1998, pp. 2791-2796.
Altman Dan W.
Lui Charlton E.
Williamson Peter H.
Chang Jon
Law Offices of Albert S. Michalik PLLC
Microsft Corporation
LandOfFree
Method and system of matching ink processor and recognizer... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system of matching ink processor and recognizer..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system of matching ink processor and recognizer... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3349580