Image analysis – Pattern recognition – Feature extraction
Reexamination Certificate
2000-01-05
2004-01-20
Chang, Jon (Department: 2623)
Image analysis
Pattern recognition
Feature extraction
C382S286000
Reexamination Certificate
active
06681046
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
This invention relates to an image processing method and, more particularly, to a scheme of specifying a plurality of cells which are set to a same attribute as a ledger slip of an inputted bit map image in setting up a character recognition program.
2. Prior Art
In work of defining a ledger slip of an OCR which causes a character recognition program to recognize a ledger slip to be newly recognized by the OCR, the work conventionally involves setting up what kind of characters, numerals and symbols, etc., are inputted in what position. By changing a recognition scheme and a set of reference samples in such work, an erroneous recognition can be avoided such as recognizing a character “O” as a numeral “O” in an area which is set up as a numeral. In addition, because the size of a set is small, a high speed recognition can be advantageously available. Because there is a character which is difficult to recognize especially in a hand written character recognition, such information is indispensable to maintain a predetermined recognition rate.
In a ledger slip definition work of an OCR, it was conventionally required to measure the size of a ledger and relative positions of page mark/position correction mark (also referred to as a timing mark or a reference mark)/character field/character using a scale and input the values thereof so that the work was very cumbersome for an operator. Particularly in a drop out ledger slip (a slip specifically designed for an OCR in which characters and a character frame are printed by a color which can not be recognized by an OCR), this method is widely employed still at the time of filing this patent application and it was necessary to consume a tremendous time to design an OCR program using a new drop out ledger slip.
In order to solve such problems in the prior art, some technologies are currently proposed. An example is a scheme describe in “WinReader Hand v1.5, Users Manual, Setting of Fields” (p19,20) (“WinReader Hand” is a trademark of Media Drive Corporation) in which an image of an unfilled ledger slip having a black frames is read in by a scanner for displaying on a display, all black vertical and horizontal lines are automatically detected, and a plurality of cells which are the units of recognition are specified to set up an attribute after an operator confirms that a straight line is correctly detected.
The above scheme requires less operator work and is much simpler in defining the ledger slip than the scheme using a scale. However, in this character recognition program, all recognition unit cells (columns) included in a ledger slip image have to be specified by an operation of clicking the upper left corner of the cell with a mouse and dragging it to and dropping at the lower right corner.
On the other hand, a multi-entry ledger slip are currently used widely in which the ledger comprises a plurality of series cells (columns) in which an identical attribute is registered as shown in FIG.
26
. In defining such multi-entry ledger slip, the above described improved scheme of ledger slip definition requires to specify each cell or each cell string and repeat a cumbersome work of registering the attribute.
SUMMARY OF THE INVENTION
It is an object of this invention to provide a system which allows a ledger slip definition work of an operator to be reduced as much as possible.
It is another object of this invention to provide a system which allows an operator to intuitively recognize a same kind of cells or cell strings to prevent an error in defining a ledger slip.
It is a further object of this invention to provide a system which prevents an error in defining a ledger slip by eliminating repetition of a simple work done by an operator.
In the present invention, image data is analyzed to detect a cell (frame in which a character or a symbol is entered) included in the image data. It is then determined whether the cells continue in a constant pattern and it is determined that the group of the continuous cells is a cell group in which a common attribute is to be registered when the cells are continuous. The cell group in which a common attribute is registered is displayed on a display device distinctively from other group of cells. The operator is allowed to modify such as delete or add a part of cells from or to specific cell group. The operator is also allowed to specify a cell group and input an attribute to all cells included in that cell group in a bundle.
This invention provides, in one aspect thereof, a method of analyzing image data inputted in an image data analyzing apparatus, said method comprising the steps of;
(a) scanning said image data to extract a plurality of cell data items including a starting coordinate value (x,y) and height information,
(b) extracting a cell group having a same x coordinate value from said plurality of cell data items,
(c) examining the height information of cells included in said extracting cell group, and
(d) eliminating a cell having a height information different from other cells included in said extracted cell group from said extracted cell group.
The term “starting coordinate value” as used herein is a concept including not only the upper left point of a rectangle circumscribing a cell but also information necessary to specify the position of a cell including upper right, lower left and lower right points of a rectangle circumscribing the cell. The term “height information” is a concept including distance information between 2 points and information of the coordinate values of 2 points in one dimension. Further, the term “cell” as used in the claims of this specification is a concept including the “cell string” to be described in the preferred embodiment of this invention.
This invention provides, in another aspect thereof, a method of analyzing image data inputted in an image data analyzing apparatus, said method comprising the steps of;
(a) scanning said image data to extract a plurality of cell data items including a starting coordinate value (x, y) and height and width information,
(b) extracting a cell group having a same x coordinate value from said plurality of cell data items,
(c) analyzing width information of cells included in said extracted cell group to extract a first cell group and a second cell group from said extracted cell group,
(d) examining the height information of cells included in said first cell group, and
(e) eliminating a cell having a height information different from other cells included in said first cell group from said first cell group.
The term “width information” as used herein is a concept including distance information between 2 points and information of the coordinate values of 2 points in a dimension other than the height.
This invention provides, in another aspect thereof, a method of analyzing image data inputted in an image data analyzing apparatus, said method comprising the steps of;
(a) scanning said image data to extract a plurality of cell data items including a starting coordinate value (x,y) and width information,
(b) extracting a cell group having a same x coordinate value from said plurality of cell data items,
(c) examining the width information of cells included in said extracted cell group, and
(d) extracting cells having substantially same width information among cells included in said extracted cell group as a new cell group.
This invention provides, in another aspect thereof, a method of analyzing image data inputted in an image data analyzing apparatus which is provided with an input device and a display device, said method comprising the steps of;
(a) scanning said image data to extract a plurality of cell data items including a starting coordinate value (x,y) and height information,
(b) extracting a cell group having a same x coordinate value from said plurality of cell data items,
(c) examining the height information of cells included in said extracted cell group, and
(d) eliminating a cell having a height information different from other cells included in said extracted cell group from sai
Fujino Yoshito
Kashiwagi Koichi
Chang Jon
Herzberg Louis P.
Scully Scott Murphy & Presser
LandOfFree
Method and apparatus for analyzing image data, storage... 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 analyzing image data, storage..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for analyzing image data, storage... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3260415