Image analysis – Pattern recognition – Unconstrained handwriting
Reexamination Certificate
1998-11-03
2001-11-20
Au, Amelia M. (Department: 2623)
Image analysis
Pattern recognition
Unconstrained handwriting
C382S229000
Reexamination Certificate
active
06320983
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method and apparatus for character recognition applied to a slip processor used in, for example, banks for determining an item selected by a customer from a slip according to a character recognition thereof, and to a computer-readable recognizing medium with a program making a computer execute the method of character recognition recorded therein.
BACKGROUND OF THE INVENTION
In recent years, there has been developed an image data reader such as an optical character reader (OCR device) for reading character information as image data (image) and then recognizing characters thereof, and efforts have been made to increase the efficiency of work by making use of this image data reader in various industrial fields.
For example, an operator attending at the window in a bank or the like can process jobs more efficiently by efficiently processing the slips with such an image data reader.
Especially, in order to achieve higher efficiency in jobs such as processing of slips as described above, it is required not only to process a large number of slips (hereafter a slip means the one dedicated to processing for character recognition using an image reader)of the same type, but also to automatically process slips having various types of formats.
Hereafter description is made for processing various types of slips using an image data reader with reference to
FIGS. 15
,
16
, and
17
.
FIG. 15
is a simulated view showing system configuration of an ordinary slip processor,
FIG. 16
is a view for explaining an example of operations in the slip processor shown in
FIG. 15
, and
FIG. 17
is a view showing a modification of the operations in the slip processor shown in FIG.
15
.
A slip processor
1000
comprises, as shown in
FIG. 15
, an image data reader
1001
for reading image data from a slip, an electronic computer
1002
for computing, and a hard disk
1003
for writing and reading data therein and therefrom. The image data reader
1001
is connected to the electronic computer
1002
and performs an operation for reading image data under controls provided by the electronic computer
1002
. It should be noted that, for example, an image scanner or a facsimile machine may be used as this image data reader
1001
.
Electronic computer
1002
as a control unit for controlling the image data reader
1001
comprises an input section
1002
-
1
such as a keyboard or a mouse for inputting data or the like (only a keyboard is shown in FIG.
15
), a main unit of the electronic computer
1002
-
2
and a display
1002
-
3
for displaying thereon data or control information or the like. Processing for character recognition of image data read by the image data reader
1001
is performed by the main unit
1002
-
2
of the computer.
The hard disc
1003
is connected to the electronic computer
1002
and, as shown in
FIG. 16
, stores therein positional information for character data, character types and information for specifying a number of digits or the like (described as define deck information hereinafter) which are to previously be recognized for each of slip types
1003
A,
1003
B,
1003
C, . . .
An image data reading/recognizing device
1005
as shown in FIG.
16
and
FIG. 17
may also be used in place of the image data reader
1001
. This image data reading/recognizing device
1005
can perform an operation for reading image data similarly to the image data reader
1001
and also execute character recognition.
Next, description is made for a case where character data described on an electricity-charge payment notice
1004
is recognized, as shown in
FIG. 16
, by using the image data reading/recognizing device
1005
. In this case, at first, a slip type (define deck information
1003
B correlated to the electricity-charge payment notice ‘slip B’
1004
in this case) is specified through an operation of the input section (hereafter keyboard)
1002
-
1
by an operator.
Then, the electronic computer
1002
takes out the define deck information
1003
B for the specified slip from the hard disk
1003
and notifies the define deck information to the image data reading/recognizing device
1005
.
With those operations, the image data reading/recognizing device
1005
can perform reading of image data and character recognition thereof according to the define deck information
1003
B as control information from the electronic computer
1002
.
However, in the technique for processing the slip as shown in
FIG. 16
, It is required for an operator to specify define deck information for each slip to be read. In this technique, a heavy work load is put on the operator, and when the amount of define deck information is large the operator may make a mistake while making an instruction. Furthermore, when thousands of types of slips are required to be processed, it is practically impossible for an operator to give instructions for the processing.
For this reason, there has been proposed a method in which slips are automatically read by previously describing ID numbers (“0102” in this case) each for previously discriminating a slip different for each slip from others at a predetermined position
1004
a
on the slip
1004
without requiring the operator to give any instruction for a slip type.
With this method, when image data for a slip is to be read by the image data reading/recognizing device
1005
, an ID number described at the predetermined position
1004
a
is first recognized, and define deck information (
1003
B in this case) correlated to the recognized ID number is used, thus character recognition being enabled.
When image data is read, however, sometimes the position at which a slip is set may be displaced from a desired position on an optical reader (e.g. a slip processor) such as the image data reader
1001
and image data reading/recognizing device
1005
. In this case, for example, even when same slip as the slip with define deck information having previously been stored in the hard disk
1003
is read out, it may be determined in the slip processor that the slip is not the one having the same layout because coordinates from a reference point (physical datum point) of the image for a character data area or a graphical area or some other area are not coincident with those in the define deck information.
In the slip processor, character recognition is not performed correctly unless the layout of read image data matches with the layout of the define deck information before execution of character recognition for the image data. And for this reason, at first, a reference point of each image is extracted, comparison is made between coordinates from the reference points of the images, and then determination is made as to whether the layouts are coincident with each other or not.
As described above, when a slip is to be read by an image scanner or the like based on the conventional technology, the direction in which the slip is rotated is specified by an operator, and image data for the slip is rotated in that specified direction, and then various types of processing such as character recognition are carried out. Alternately, there is employed a method for any preset slip that particular information such as a correction mark on the slip is used and the slip is rotated in the correct direction.
When a slip is an application form for money transfer used In a bank, and if processing for character recognition is executed to the image data of a slip for which the reading direction has been changed, a type of account to be checked in relation to a method of transferring money into any other account can be identified. An image reading mechanism based on recognition by an OCR (Optical Character Reader) is generally provided for the purpose described above. Therefore, an OCR-dedicated form is used as an application form for money transfer, on which boxes for marking or character strings each for specifying an account type are printed. Those previously printed boxes or the character strings are dropped out when the image is read out, so that only the marks wr
Chiba Koichi
Eguchi Shinichi
Kobara Katsutoshi
Matsuno Hideki
Nagano Yoshihiro
Arent Fox Kintner & Plotkin & Kahn, PLLC
Au Amelia M.
Dastouri Mehrdad
Fujitsu Limited
LandOfFree
Method and apparatus for character recognition, and... 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 apparatus for character recognition, and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for character recognition, and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2586112