JP4164976B2 - Character recognition device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for designating a recognition target area in a character recognition device.
[0002]
[Prior art]
Conventionally, as a technique in such a field, for example, there are those described in the following documents.
Document: Japanese Patent No. 2661898 Specification
FIG. 2 is an explanatory diagram of a method for specifying a recognition target area in the conventional character recognition apparatus described in the above-mentioned document.
In this character recognition device, when specifying a recognition target area Q surrounded by a ruled line on an image displayed on the display unit, the cursor is moved to an arbitrary point q in the area Q and specified. The method is taken. When the coordinates of one point q are input by the cursor, a ruled line search is performed in four directions, up, down, left, and right, centering on this point q. Then, the area surrounded by the two horizontal ruled lines detected first in the vertical search and the two vertical ruled lines detected first in the horizontal search is specified as the designated recognition target area Q. It was supposed to be.
[0004]
[Problems to be solved by the invention]
However, the conventional character recognition device has a problem that, for example, two adjacent areas each surrounded by a ruled line cannot be designated as one recognition target area.
The present invention solves the problems of the prior art and provides a character recognition device that can designate a plurality of areas each surrounded by a ruled line as one recognition target area.
[0005]
[Means for Solving the Problems]
In order to solve the above-described problems, the present invention provides a character recognition apparatus, wherein an image input unit that inputs a document in which a character to be recognized is described in a plurality of areas divided by ruled lines as an image image, and the image input unit A region designating unit that designates the recognition target region on the image image input in the first region and the second point in the same region or different regions divided by the ruled line, and the following region determination unit, A character cutout unit and a character recognition unit are provided.
[0006]
The area determination unit searches the opposite side of the second point from the first point in the horizontal direction and the vertical direction, extracts the first vertical ruled line and the horizontal ruled line, and extracts the second point. As a starting point, a second vertical ruled line and a horizontal ruled line are extracted by searching the opposite side of the first point in the horizontal and vertical directions, and a range surrounded by the first and second vertical ruled lines and horizontal ruled lines. Is determined as the recognition target area. The character cutout unit cuts out a character image to be recognized from the recognition target region determined by the region determination unit, and the character recognition unit compares the character image with a character dictionary and applies a corresponding character code. Is output.
[0007]
According to the present invention, since the character recognition device is configured as described above, the following operation is performed.
When the first and second points are designated by the region designating unit on the image image of the document input by the image input unit, the region determining unit searches for the image image starting from the first point, First vertical ruled lines and horizontal ruled lines are extracted. Further, the image image is searched starting from the second point, and the second vertical ruled line and horizontal ruled line are extracted. A range surrounded by the extracted first and second vertical ruled lines and horizontal ruled lines is determined as a recognition target area and given to the character cutout unit. In the character cutout unit, a character image to be recognized is cut out based on the given recognition target region and output to the character recognition unit. In the character recognition unit, the given character image is checked against the character dictionary, and the corresponding character code is selected and output.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
(First embodiment)
FIG. 1 is a block diagram of a character recognition apparatus showing a first embodiment of the present invention.
This character recognition device includes an image input unit 1 such as an image scanner that separates a ruled line for dividing a writing surface into a plurality of areas and a document in which characters are written in these areas into pixels and reads them as image images. I have. Connected to the output side of the image input unit 1 is an image memory 2 for storing the read image image as, for example, monochrome binary pixel data. The image memory 2 is connected to a display unit 3 that displays an image image stored in the image memory 2 and displays information for specifying an area or the like superimposed on the image image.
[0009]
The character recognition apparatus also includes an area designating unit 4 and an area determining unit 5 for specifying a recognition target area. The area specifying unit 4 has an input device such as a keyboard and a mouse, and specifies two points (for example, points a and b) for specifying a recognition target area on the image displayed on the display unit 3. Input. The output side of the area specifying unit 4 is connected to the area determining unit 5. The area determination unit 5 searches the image memory 2 based on the coordinates of the points a and b designated by the area designation unit 4 and extracts ruled lines that are boundaries of the recognition target area Z. That is, the first vertical ruled line and the horizontal ruled line are extracted by searching the opposite side of the point b from the point a, and the second vertical ruled line and the horizontal ruled line are searched by searching the opposite side of the point a from the point b. Is supposed to be extracted. The area determination unit 5 determines the area surrounded by the first and second vertical ruled lines and horizontal ruled lines as the recognition target area Z.
[0010]
The area determination unit 5 is connected to a character extraction unit 6 that extracts a character image from the determined recognition target area Z. Connected to the output side of the character cutout unit 6 is a character recognition unit 8 for recognizing characters according to the feature data of the cutout character image with reference to the character dictionary 7 storing the data of the patterned characters. Has been. A character code of a candidate character as a recognition result is output from the character recognition unit 8.
[0011]
FIG. 3 is an explanatory diagram of processing in the area designating unit 4 and the area determining unit 5 in FIG. In FIG. 3, the recognition target image stored in the image memory 2 and displayed on the display unit 3, the points a and b specified by the region specifying unit 4, and the recognition target region Z specified by these two points are shown. Is illustrated. Hereinafter, the operation of FIG. 1 will be described with reference to FIG. 3, focusing on the recognition target area specifying process, which is a feature of the present invention.
When the document to be read is set in the image input unit 1 and read, the read image is stored in the image memory 2 as monochrome binary pixel data. The image image stored in the image memory 2 is further displayed on the display unit 3.
[0012]
An operator performs input using an input device in order to designate a recognition target area. When using, for example, a mouse as an input device of the area designating unit 4, a mouse cursor is displayed on the recognition target area image of the display unit 3 as shown in FIG. 3, and two points a and b are clicked and designated. To do. The coordinates a (Xa, Ya) of the point a thus designated and the coordinates b (Xb, Yb) of the point b are given from the area designating unit 4 to the area determining unit 5.
[0013]
In the region determination unit 5, the recognition target region Z is determined by the following procedure based on the coordinates of the two points a and b and the image image stored in the image memory 2.
First, the point on the left side is determined by comparing the X coordinate values of the points a and b. Here, it is assumed that the point a is on the left side. Further, the point on the upper side is determined by comparing the values of the Y coordinates of the points a and b. Here, it is assumed that the point b is on the upper side.
[0014]
Next, starting from the point a on the left side, the first vertical ruled line LL on the left side is searched. In the vertical ruled line search, a part where a predetermined number of black pixels continue in the vertical direction is detected as a ruled line. Similarly, starting from the point b on the right side, the first vertical ruled line LR on the right side is searched. Also, starting from the point b on the upper side, the first horizontal ruled line LU on the upper side is searched. In the horizontal ruled line search, a part where a predetermined number of black pixels continue in the horizontal direction is detected as a ruled line. Similarly, starting from the lower point a, the first horizontal ruled line LD on the lower side is searched.
[0015]
Then, a rectangular range surrounded by the two vertical ruled lines LL and LR and the two horizontal ruled lines LU and LD is determined as the recognition target region Z.
Information on the recognition target region Z determined in this way is given from the region determination unit 5 to the character cutout unit 6. The character cutout unit 6 sequentially cuts out the character images in the recognition target area Z from the image memory 2 and gives them to the character recognition unit 8. In the character recognition unit 8, feature data of the given character image is extracted, the character dictionary 7 is referred to, and the character code of the corresponding candidate character is output.
As described above, the character recognition device according to the first embodiment determines, as the recognition target region Z, the range surrounded by the ruled line including these two points a and b according to the designated two points a and b. A region specifying unit 4 and a region determining unit 5 are provided. Accordingly, there is an advantage that even two or more areas partitioned by ruled lines can be designated as one continuous recognition target area Z.
[0017]
(Second Embodiment)
FIG. 4 is a block diagram of a character recognition apparatus showing a second embodiment of the present invention. Elements common to those in FIG. 1 are denoted by common reference numerals.
In the character recognition device of this embodiment, switching units 9 and 10 are provided between the image memory 2 and the region determination unit 5 and between the region determination unit 5 and the character cutout unit 6 of the character recognition device of FIG. The second region designating unit 11 and the region determining unit 12 are provided between the switching units 9 and 10. Further, a cursor storage unit 13 for the display unit 3 is added.
[0018]
The switching units 9 and 10 select one of the region determination units 5 and 12 according to the selection signal SEL. Further, the second region specifying unit 11 and the region determining unit 12 determine the recognition target region by, for example, the conventional method shown in FIG. Further, the cursor storage unit 13 stores the scale shape and color of each cursor displayed on the display unit 3 in accordance with the recognition target region determination method of the two region determination units 5 and 12, and selects the selection signal. The cursor display corresponding to the method selected by SEL is performed. Other configurations are the same as those in FIG.
[0019]
In such a character recognition device, when the region determination unit 5 is selected by the selection signal SEL, the same operation as in the first embodiment is performed. Further, when the region determining unit 12 is selected by the selection signal SEL, the same operation as the conventional one is performed. At this time, the cursor display displayed on the display unit 3 is different corresponding to the region determination unit (5 or 12) selected by the selection signal SEL, and it is clear which method is used to specify and determine the recognition region. Are displayed separately.
[0020]
As described above, since the character recognition device according to the second embodiment includes the second region designation unit 11 and the region determination unit 12, two types of recognition target region designation methods can be switched and used. It becomes possible. Furthermore, since the cursor storage unit 13 for switching the cursor display corresponding to the selected method is provided, a suitable method can be definitely selected when the recognition target area is designated.
[0021]
In addition, this invention is not limited to the said embodiment, A various deformation | transformation is possible. Examples of such modifications include the following (a) to (d).
(A) In FIG. 1, each unit such as the region specifying unit 4 and the region determining unit 5 is configured as an individual processing unit, but may be processed by program control using a microprocessor or the like.
(B) In FIG. 3, the processing in the case where the two points a and b are designated in different areas has been described, but two points in the same area may be designated.
[0022]
(C) In FIG. 3, the case where three regions arranged in the horizontal direction are designated as the recognition target region Z is illustrated, but m × n regions arranged in the horizontal direction and the vertical direction are designated as the recognition target region. You can also In that case, for example, it is only necessary to designate two points: a point a in the upper left area and a point b in the lower right area.
(D) Although the image is processed as black and white, it can be similarly applied to a color image.
[0023]
【The invention's effect】
As described above in detail, according to the present invention, an area designating unit that can designate two points in different areas and a ruled line surrounding the designated two points are extracted in order to identify the recognition target area. And an area determining unit that determines the recognition target area. Thus, a plurality of areas each divided by a ruled line can be designated as one recognition target area.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a character recognition device according to a first embodiment of the present invention.
FIG. 2 is an explanatory diagram of a method for specifying a recognition target area in a conventional character recognition apparatus.
FIG. 3 is an explanatory diagram of processing in an area designating unit 4 and an area determining unit 5 in FIG.
FIG. 4 is a block diagram of a character recognition device showing a second embodiment of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Image input part 2 Image memory 3 Display part 4 Area designation part 5 Area determination part 6 Character extraction part 7 Character dictionary 8 Character recognition part

Claims (1)

An image input unit for inputting, as an image image, a document in which characters to be recognized are described in a plurality of areas divided by ruled lines;
A region designating unit for designating a recognition target region on the image image input by the image input unit by the first and second points in the same region or different regions separated by the ruled lines;
A first vertical ruled line and a horizontal ruled line are extracted by searching the opposite side of the second point in the horizontal direction and the vertical direction from the first point, and the first point is used as the first point. The second vertical ruled line and the horizontal ruled line are extracted by setting the opposite side of the point to the horizontal and vertical directions, and the range surrounded by the first and second vertical ruled lines and horizontal ruled lines is the recognition target area. An area determination unit to determine as
A character cutout unit that cuts out a character image to be recognized from the recognition target region determined by the region determination unit;
A character recognition unit that collates the character image with a character dictionary and outputs a corresponding character code;
A character recognition device comprising:

Images