Image analysis – Pattern recognition – Classification
Reexamination Certificate
1998-07-31
2001-11-20
Boudreau, Leo (Department: 2721)
Image analysis
Pattern recognition
Classification
C382S159000, C382S181000, C382S186000, C382S187000
Reexamination Certificate
active
06320985
ABSTRACT:
BACKGROUND
1. Technical Field
The present application relates generally to data classification and, more particularly, to an apparatus and method for improving recognition accuracy in handwritten text recognition systems by augmenting character data of collected handwriting samples.
2. Description of the Related Art
Currently, the need for accurate machine recognition of handwritten text has increased due to the popularity and wide spread use of handheld, pen-based computers. However, achieving high accuracy for handwriting recognition in conventional devices has proven to be a difficult task due to the wide variety of handwriting styles, many of which have ambiguous and/or conflicting character representations. In order to combat this problem, various techniques have been developed to enable handwriting recognition devices to adapt to an individual's writing style. These conventional methods can generally be divided into two categories: methods which require the collection of handwriting samples from each person; and methods which do not require such samples.
Typically, the conventional recognition systems which utilize handwriting samples are preferably used due to their superior recognition performance. However, one inherent problem with collecting these handwriting samples is that it is typically a tedious and burdensome process for the writer. In order to mitigate this burden and encourage the collection of the required samples, the collection process can be made easier by making the amount of required writing samples as small as reasonably possible. Reducing the amount of required writing samples, however, leads to another problem: the probability increases that samples of individual characters will be omitted. Consequently, when characters are omitted from the collected handwriting samples, the ability to achieve accurate handwriting recognition diminishes.
SUMMARY
The present application is generally directed to an apparatus and method for providing improved data classification and, in particular, to an apparatus and method for improved handwriting data classification which enables handwriting recognition devices to robustly handle and recover from the problems associated with the omission of characters from collected handwriting samples.
In one aspect, a data classification apparatus comprises:
means for inputting a plurality of data, the plurality of data including one of data to be recognized, generic data and user-specific data;
means for augmenting the user-specific data with the generic data to generate augmented user-specific data;
means for training the data classification apparatus with the augmented user-specific data to generate training data; and
means for recognizing the data to be recognized in accordance with the training data.
In another aspect, in a data classification apparatus having means for inputting a plurality of data, the plurality of data including at least one of data to be recognized and user-specific data, the apparatus further having means for generating training data from the user-specific data, and means for recognizing the data to be recognized with the training data, a method for augmenting the user-specific data to improve classification accuracy of the apparatus, comprising the steps of:
collecting a plurality of generic user data;
computing threshold data in accordance with the collected plurality of generic user data;
augmenting the user-specific data with the generic user data in accordance with the computed threshold data, whereby the augmented user-specific data is used by the training means to generate the training data.
REFERENCES:
patent: 5005205 (1991-04-01), Ellozy et al.
patent: 5058182 (1991-10-01), Kuan et al.
patent: 5333209 (1994-07-01), Sinden et al.
patent: 5434929 (1995-07-01), Beernink et al.
patent: 5454046 (1995-09-01), Carman, II
patent: 5479523 (1995-12-01), Gaborski et al.
patent: 5559897 (1996-09-01), Brown et al.
patent: 5559925 (1996-09-01), Austin
patent: 5577135 (1996-11-01), Grajski et al.
patent: 5617486 (1997-04-01), Chow et al.
patent: 5642435 (1997-06-01), Loris
patent: 5644652 (1997-07-01), Bellegarda et al.
patent: 5666438 (1997-09-01), Beernink et al.
patent: 5675665 (1997-10-01), Lyon
patent: 5680480 (1997-10-01), Beernink et al.
patent: 5680510 (1997-10-01), Hon et al.
patent: 5689584 (1997-11-01), Kobayashi
patent: 5710831 (1998-01-01), Beernink et al.
patent: 5737487 (1998-04-01), Bellegarda et al.
patent: 5737490 (1998-04-01), Austin et al.
patent: 5757964 (1998-05-01), Lee et al.
patent: 5761687 (1998-06-01), Hon et al.
patent: 5768422 (1998-06-01), Yaeger
patent: 5796863 (1998-08-01), Lyon
patent: 5805730 (1998-09-01), Yaeger et al.
patent: 5805731 (1998-09-01), Yaeger et al.
patent: 5828999 (1998-10-01), Bellegarda et al.
patent: 5852801 (1998-12-01), Hon et al.
patent: 5903884 (1999-05-01), Lyon et al.
patent: 5940535 (1999-08-01), Huang
patent: 5946410 (1999-08-01), Lyon
patent: 6052481 (2000-04-01), Grajski et al.
Exhibit A of Supplemental Declaration of Michael Perrone, dated Aug. 25, 2000, program code for traugdat. c.
Declaration of Michael P. Perrone.
Perrone Michael P.
Subrahmonia Jayashree
Boudreau Leo
F. Chau & Associates LLP
International Business Machines - Corporation
Mariam Daniel G.
LandOfFree
Apparatus and method for augmenting data in handwriting... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for augmenting data in handwriting..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for augmenting data in handwriting... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2589644