JPH11134435A - Method for normalizing character picture and device for recognizing character - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a character recognition device equipped with a normalization processing means for attaining stable recognition without necessitating plural dictionaries even when the thickness of a line in one character fluctionates. SOLUTION: Thinning processing of a character picture for one character segmented by character segmenting part 110 is operated by a thinning processing part 121, and a circumscribed rectangle of the character is detected by a circumscribed rectangle detecting part 122 for the thinned character picture, and the enlargement or reduction judgement of the picture is operated according to the size of the circumscribed rectangle. When the enlargement judgement is operated, the enlargement normalization processing of the thinned character picture is operated by an enlargement normalizing part 123 so that a picture arth a normalized size can be obtained. When the reduction judgement is operated, the OR normalization processing of the thinned character picture is operated by an OR normalizing part 124 so that a picture with a normalized size can be obtained. Then, the picture whose size is normalized by the enlargement normalizing part 123 or the OR normalizing part 124 is processed by a thickness processing part 125 so that the width of the segment line of the picture can be turned into preliminarily set thickness.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device and, more particularly, to a technique for normalizing character images having different line widths and character sizes.

[0002]

2. Description of the Related Art In a character recognition apparatus, the line width of a read character image is not uniform, the character size is not constant, and character images having various line thicknesses and sizes can be obtained. To perform character recognition from such a character image, the read character image is normalized as preprocessing.

As methods for normalizing the thickness and size of lines of a read character image, a thinning normalization technique and an OR normalization technique are known. Here, the thinning-out normalization is a method of extracting a pixel at a fixed interval and generating a pixel smaller in size than the input pixel by truncating other pixels. This is a method of generating an image having a smaller size than the input image by taking (logical sum).

[0004] However, thinning normalization has a problem that fine lines are easily lost or faded, and pixels are liable to collapse in a character loop in OR normalization. There was a problem.

In consideration of the above problems, a method of performing line segment thinning processing on an input image and performing OR normalization on the image after the thinning processing (Japanese Patent Laid-Open No. 3-3778).
No. 3).

Further, as an improvement of the technique disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 3-37783, a line segment thinning process is performed on an input image while changing its strength, and each character image after the thinning process is applied to the input image. By performing OR normalization on the image, a plurality of normalized character images are obtained for one image, each of them is subjected to recognition processing, and a final recognition result is obtained from the obtained plurality of recognition results (Japanese Patent Laid-Open No. 4-687). No.).

[0007]

SUMMARY OF THE INVENTION The above-mentioned Japanese Patent Application Laid-Open No. 4-687.
In the character recognition technique disclosed in Japanese Patent Application Laid-Open No. H11-260, the normalization of line segments applied to an input image is changed, and OR normalization is performed on each character image after the thinning process. If the degree of change in thickness is large, the number of narrowing processes increases in accordance with the degree, the number of characters normalized by OR normalization also increases, and the number of recognition processes increases. There is a certain limit in terms of memory capacity and recognition processing speed, for example, when the change in the width of is larger than a certain value, the processing speed becomes slower.

For example, in the case of a character written with a ballpoint pen, for example, ink may accumulate on the pen tip of the ballpoint pen, and if writing is continued as it is, the writing line becomes thicker or thinner in one character. In some cases, the thickness of the line segment in the character changes greatly. Further, in the case of a character enlarged by copying or the like, the line width becomes thicker when the magnification is large, and the image quality becomes coarse.

There are various types of writing implements such as pencils, ball-point pens, and felt-tip pens, and there are various types of core thicknesses.
As described above, when a character having all line widths of a character written by a writing instrument is subjected to recognition processing by the same pattern matching, a large number of dictionaries corresponding to the line width are required by using a feature extraction method in which feature data changes depending on the line width. That means
There is a problem that the efficiency of the recognition processing is reduced.

The present invention has been made in view of the above-mentioned problems of the prior art, and enables stable recognition without the need of a dictionary corresponding to the line width even if the thickness of a line in one character varies. It is an object of the present invention to provide a character recognition device including a character image normalization method and a normalization processing unit that can be used.

[0011]

In order to solve the above-mentioned problems, a character image normalization method according to a first aspect of the present invention extracts a character image from an input image, performs a normalization process, and then performs normalization. In a character recognition device that performs character recognition processing on a character image, a step of performing thinning processing on the cut-out character image, detecting a circumscribed rectangle of the character image after the thinning processing, and determining the size of the A step of determining whether to enlarge or reduce the image by normalizing; a step of performing an enlargement normalization process or a reduction normalization process on the character image based on the determination result; Performing a thickening process on the obtained character image to obtain a character image having a line width of a predetermined thickness.

According to a second aspect of the present invention, there is provided a character recognition method for extracting a character image from an input image, performing a normalization process, and then performing a character recognition process on the normalized character image. A step of detecting a circumscribed rectangle of the cut-out character image and determining whether to enlarge or reduce the character image based on its size; and Performing thinning processing or reduction normalization processing, performing thinning processing on a character image that has been subjected to enlargement normalization processing or reduction normalization processing, and performing thickening processing on the character image after thinning processing and performing predetermined processing. And forming a character image having a line width of the thickness of the character image.

According to a third aspect of the present invention, there is provided a character recognition method for extracting a character image from an input image, performing a normalization process, and performing a character recognition process on the normalized character image. In the apparatus, a process of detecting a circumscribed rectangle of the cut-out character image and performing only enlargement normalization on the cut-out character image based on the size of the circumscribed rectangle and the ratio of the number of pixels of the character occupying the circumscribed rectangle Or performing thinning normalization, a step of determining whether to perform thinning, and a step of performing enlargement normalization processing, thinning normalization processing, or thinning processing of the character image based on the determination result, Further, after the step of thinning processing, a step of detecting a circumscribed rectangle of the character image after thinning processing and determining whether to enlarge or reduce the character image based on its size, Based on the result A process of performing an enlargement normalization process or a reduction normalization process on the linearized character image; and performing a thickening process on the character image that has been subjected to the enlargement normalization process or the reduction normalization process to obtain a line width of a predetermined thickness. Forming a character image.

According to a fourth aspect of the present invention, there is provided a character recognition method for extracting a character image from an input image, performing a normalization process, and then performing a character recognition process on the normalized character image. In the apparatus, a process of detecting a circumscribed rectangle of the cut-out character image and performing only enlargement normalization on the cut-out character image based on the size of the circumscribed rectangle and the ratio of the number of pixels of the character occupying the circumscribed rectangle Deciding whether to perform only normalization or thinning normalization, or to perform enlargement normalization, thinning normalization, or reduction, and based on the judgment result, enlargement normalization processing, thinning normalization processing, or Performing a reduction normalization process; and further comprising, when the determination result is expansion normalization, thinning normalization, or reduction, corresponding expansion normalization processing, thinning normalization processing,
Or a step of performing reduction normalization processing, and a step of subsequently performing thinning processing on a character image that has been subjected to enlargement normalization processing, thinning normalization processing, or reduction normalization processing, and a step of performing the thinning character image. Converting the thickening process to a character image having a line width of a predetermined thickness.

The character recognition device according to a fifth aspect of the present invention cuts out a character image from an input image and performs a normalization process.
In a character recognition device that performs character recognition processing on a normalized character image, a thinning processing unit that performs thinning processing on the cut-out character image and a circumscribed rectangle of the character image after the thinning processing are detected. Circumscribed rectangle detecting means for determining whether to enlarge or reduce the character image in accordance with the size of the character image, and enlargement normalization processing for enlarging and normalizing the character image based on the determination result of the circumscribed rectangle detection means Means, a reduction normalizing means for performing reduction normalization processing of the character image based on the determination result of the circumscribing rectangle detection means, and a thickening processing for applying a thickening processing to the output of the enlargement normalization means or reduction normalization means. And a character normalizing means comprising a thickening means for forming a character image having a line width.

The character recognition apparatus according to a sixth aspect of the present invention cuts out a character image from an input image and performs a normalization process.
In a character recognition device that performs character recognition processing on a normalized character image, whether a circumscribed rectangle of the cut-out character image is detected and whether the character image is enlarged or normalized according to its size Circumscribed rectangle detection means for determining
Enlargement normalizing means for performing enlargement normalization processing of the character image based on the determination result of the circumscribed rectangle detection means; reduction normalization means for performing reduction normalization processing of the character image based on the determination result of the circumscribed rectangle detection means; A thinning processing means for performing thinning processing on the output of the enlargement normalizing means or the reduction normalizing means; and a character image having a line width of a predetermined thickness by performing thickening processing on the output of the thinning processing means. And a character normalizing means comprising:

The character recognition device according to a seventh aspect of the present invention cuts out a character image from an input image and performs a normalization process.
In a character recognition device for performing character recognition processing on a normalized character image, a first circumscribed rectangle detecting means for detecting a circumscribed rectangle of the cut-out character image, a size of the circumscribed rectangle and a character occupying the circumscribed rectangle Character pixel detecting means for determining whether to perform only enlargement normalization, thinning normalization, or thinning of the extracted character image based on the ratio of the number of pixels Normalization processing means for performing an expansion normalization process on the character image based on the result; reduction normalization means for performing a reduction normalization process on the character image based on the determination result of the character pixel detection means; and determination by the character pixel detection means Thinning processing means for performing thinning processing of the character image based on the result, further detecting a circumscribed rectangle of the character image after the thinning processing, and enlarging and normalizing the character image according to the size. Or reduced second determines normalized
And performing a fattening process on the character image enlarged or reduced by the enlargement normalizing means or the reduced enlargement normalizing means based on the determination result of the circumscribed rectangle detection means and the second circumscribed rectangle detection means. A character normalizing unit having a thickening processing unit for forming a character image having a line width of a predetermined thickness.

The character recognition device according to an eighth aspect of the present invention cuts out a character image from an input image and performs a normalization process.
In a character recognition device that performs a character recognition process on a normalized character image, a circumscribed rectangle detection unit that detects a circumscribed rectangle of the cut-out character image, a size of the circumscribed rectangle, and the number of pixels of the character occupying the circumscribed rectangle Character pixel detection to determine whether to perform only enlargement normalization or only thinning normalization, or to perform enlargement normalization, thinning normalization, or reduction on the extracted character image based on the ratio of Means,
Enlargement normalizing means for performing enlargement normalization processing of a character image based on the determination result of the character pixel detection means, and thinning normalization means for performing thinning normalization processing of the character image based on the determination result of the character pixel detection means; Reduction normalization means for performing a reduction normalization process on the character image based on the determination result of the character pixel detection means, and further comprising a determination result of the character pixel detection means which is enlarged, thinned, or reduced. In the case of (1), thinning processing means for performing thinning processing on the character image after the enlargement normalization processing, thinning normalization processing, or reduction normalization processing, and performing thickening processing on the thinned character image by a predetermined A character normalizing unit having a thick processing unit for forming a character image having a line width of a thickness.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A character image normalizing method according to the present invention detects a circumscribed rectangle of a line segment of a cut-out character image, and enlarges, reduces, thins out, or thins the line according to the size of the circumscribed rectangle. , The width of the line segment is made substantially constant.

<Embodiment 1> First of the normalization methods of the present invention
Performs thinning processing of the line segment of the extracted character image, detects the circumscribed rectangle, performs expansion normalization or reduction normalization according to the size of the circumscribed rectangle, and thickens the obtained line segment. To make the line segment width constant.

FIG. 1 is a block diagram showing an embodiment of a character recognition apparatus to which the first normalization method is applied. The character recognition apparatus 10 includes a character cutout unit 110 and a normalization processing block 12.
0, and a character recognition unit 130. The normalization processing block 120 includes a thinning processing unit 121, a circumscribed rectangle detection unit 122, an expansion normalization unit 123, an OR normalization unit 124, and a thick processing unit 125. I have.

The character extracting section 110 extracts character images one by one from the image Im read by a character reading device such as a scanner.

The normalization processing block 120 detects the circumscribed rectangle after performing the thinning process of the line of the extracted image of one character, and performs enlargement normalization or reduction normalization according to the size of the circumscribed rectangle. The size of the line is made constant, and the width of the obtained line segment is set to a predetermined thickness and given to the recognition processing unit 130. For this purpose, first, the thinning processing section 121 performs thinning processing on the character image of one character cut out by the character cutout section 110, and obtains the thinned image. Next, the circumscribed rectangle detection unit 122 detects the circumscribed rectangle of the character from the image thinned by the thinning processing unit 121, and determines the enlargement or reduction of the character image based on the size of the circumscribed rectangle. That is, in the circumscribed rectangle detection unit 122, when the size of the circumscribed rectangle is n × m dots and n ′ × m ′ is the normalized size of the character image, n and m are smaller than n ′, and n and m are smaller than n ′. When it is smaller than m ′, it is determined that the character image needs to be enlarged, and when at least one of n and m is larger than n ′ or n, it is determined that the character image needs to be reduced.

If the circumscribed rectangle detection unit 122 determines that the enlargement has been performed, the enlargement normalization unit 123 performs the thinning processing unit 122.
The thinned character image is enlarged and normalized to obtain an image whose size is normalized. If the reduction is determined, the OR normalization unit 124 sets the thinning processing unit 12
The character image thinned by 2 is subjected to OR normalization processing to obtain an image whose size is normalized.

The expansion normalizing unit 123 or the OR normalizing unit 1
The thickening processing section 125 performs thickening processing on the character image whose size has been normalized at 24 so that the line segment width of the character image becomes a predetermined thickness.

The character recognition unit 130 performs character recognition based on the normalized image generated by the normalization processing block 120, and outputs the character code as a recognition result.

[Example] Normalized block 1 of the present embodiment
One embodiment of the invention will now be described. 5 shows an example of an input image, FIG. 6 shows an example of a character image cut out from the image of FIG. 5, and FIG. 7 shows an example of a character image subjected to thinning processing. 8 is a diagram illustrating an example of a character image that has been enlarged and normalized, and FIG. 9 is a diagram illustrating an example of a normalized character image. An image Im (FIG. 5) read from a scanner or the like is cut out one character at a time and has a line width of 3 dots and a size of 30 dots.
In the case of normalizing to a character image of × 30 dots (FIG. 9) as an example, in the character image of five characters cut out by the character cutout unit 120, the writing start is thick because ink is accumulated at the pen tip of the ballpoint pen. '2' with halfway
The case where the character image (FIG. 6) is normalized will be described.

The character image 61 (FIG. 6) representing the character "2" taken into the normalization block 120 is first subjected to thinning processing by the thinning processing section 121, and the thinned character image as shown in FIG. Becomes Next, a circumscribed rectangle (FIG. 7) of the character image 71 thinned by the circumscribed rectangle detection unit 122 is obtained (in the example of FIG. 7, a circumscribed rectangle 72 of 19 × 26 dots is obtained). Here, n = 19, m = 26, n ′ = 30, m ′
= 30, so that n and m are smaller than n ′ and
Since n and m are smaller than m ′, the extended normalization processing unit 12
In step 3, the character size is enlarged and normalized, and a character image 81 in which the character size is enlarged and normalized as shown in FIG. 8 is obtained. Next, thickening processing is performed on the normalized character image 81 by the thickening processing unit 125, and the thickness is reduced to 3 as shown in FIG.
A character image 91 having a dot size of 30 × 30 dots is obtained.

The circumscribed rectangle of the character image thinned by the circumscribed rectangle detection unit 122 has a predetermined size n ′ ×
If it is larger than m ′ dots (for example, 35 × 24 dots, 21 dots × 34 dots, or 36 × 40 dots), a normalized image whose size has been reduced by the OR normalization unit 124 is obtained, and The thickening process is performed by the processing unit 125, and a predetermined size of n ′ × m ′ dots (3 in this example)
A character image 91 of (0 × 30 dots) is obtained.

Thus, the normalization processing block 12
In the normalized image (FIG. 9) generated at 0, the portion where the line segment has a large change in thickness (for example, a thick portion at the beginning of writing as shown in FIG. 6 or a portion blurred in the middle) is almost the same. Line width. Therefore, stable character recognition in the character recognition unit 130 is realized without requiring a plurality of dictionaries corresponding to character widths.

<Embodiment 2> Second embodiment of the normalization method of the present invention
Is a line segment obtained by detecting a circumscribed rectangle of a cut-out character image, performing expansion normalization or reduction normalization according to the size of the circumscribed rectangle, and then performing line thinning processing of an input image. To a substantially constant width.

FIG. 2 is a block diagram of an embodiment of a character recognition apparatus to which the second normalization method is applied. The character recognition apparatus 20 includes a character cutout unit 110 and a normalization processing block 22.
0, and a character recognition unit 130. The normalization processing block 220 includes a circumscribed rectangle detection unit 221, an expansion normalization unit 222, an OR normalization unit 223, a thin line processing unit 224,
And a thick processing unit 225.

The character extracting section 110 extracts character images one by one from the image Im read by a character reading device such as a scanner (FIG. 6).

The normalization processing block 220 detects a circumscribed rectangle of the image of one character, performs enlargement normalization or reduction normalization according to the size of the circumscribed rectangle, and then extracts a line segment of the extracted image of one character. Is performed, and the width of the obtained line segment is set to a predetermined thickness and given to the recognition processing unit 130. For this purpose, first, the circumscribed rectangle detection unit 221 detects the circumscribed rectangle of the character from the character image cut out by the character cutout unit 110, and determines whether to enlarge or reduce the image based on the size of the circumscribed rectangle. . Note that the enlargement / reduction determination method is the same as the circumscribed rectangle detection unit 12 in the first embodiment.
It may be the same as 1.

If the circumscribed rectangle detection unit 221 determines enlargement, the enlargement normalization unit 222 enlarges the image by enlargement normalization processing and enlarges it to obtain a character image whose size is normalized. When the reduction is determined, the image is OR-normalized by the OR normalizing unit 223 to obtain a character image whose size is normalized.

The thinning processing unit 221 performs thinning processing on the character image whose size has been normalized by the enlargement normalizing unit 222 or the OR normalizing unit 223, and obtains the thinned image. Then, the thickening processing unit 225 sets the thickness of the character image thinned by the thinning processing unit 221 to a preset thickness.

The character recognizing section 130 performs character recognition based on the normalized image generated by the normalization processing block 220, and outputs the character code as a recognition result.

Thus, similar to the first embodiment,
In the normalized image (FIG. 9) generated in the normalization processing block 220, a portion where a change in the thickness of the line segment is large has almost the same line width. Therefore, stable character recognition in the character recognition unit 130 is realized without requiring a plurality of dictionaries corresponding to character widths.

<Embodiment 3> Third embodiment of the normalization method of the present invention
Detects the circumscribed rectangle of the extracted character image, and performs thinning normalization, enlargement normalization, or reduction normalization according to the ratio of the number of pixels of the character image to the circumscribed rectangle.
Thickness processing is performed on the character image that has been subjected to enlargement normalization or reduction normalization, so that the width of the line segment is almost constant,
The character image subjected to thinning-out normalization is output as it is. In addition, when the line width of the character is not fixed, the character recognition is not adversely affected, that is, when the circumscribed rectangle is larger than the predetermined value C, the size thinning process is performed without performing the thickening process. Also, the size of the circumscribed rectangle is a predetermined value D
Also in the case between C and C, the processing speed is increased by performing size expansion normalization processing without performing thickening processing.

FIG. 3 is a block diagram of an embodiment of a character recognition apparatus to which the third normalization method is applied. The character recognition apparatus 30 includes a character cutout unit 110 and a normalization processing block 32.
0, and a character recognition unit 130. The normalization processing block 320 includes a circumscribed rectangle detection unit 321, a character pixel detection unit 322, a thin line processing unit 323, and a circumscribed rectangle detection unit 32.
4, an expansion normalizing unit 325, an OR normalizing unit 326, a thinning processing unit 327, and a thickening processing unit 328.

The character extracting section 110 extracts character images one by one from the image Im read by a character reading device such as a scanner.

The normalization processing block 320 detects a circumscribed rectangle of the image of one character, and performs thinning normalization, enlargement normalization, or reduction normalization according to the ratio of the number of pixels of the character image to the circumscribed rectangle. After that, the character image subjected to the enlargement normalization or the reduction normalization is subjected to a thickening process, and the obtained line segment is given to the recognition processing unit 130 with a predetermined width. Note that if the character line width is not fixed and the character recognition is not adversely affected, that is, if the circumscribed rectangle is larger than the predetermined value C, the size is thinned without performing the thickening process, and the character recognition is performed. Output to the unit 130. Also, in the case where the size of the circumscribed rectangle is between the predetermined values D and C, the enlargement normalization process is performed without performing the thickening process, and the character recognition unit 130
Output to

For this purpose, the circumscribed rectangle detection unit 321 detects a circumscribed rectangle of a character in the character image of one character cut out by the character cutout unit 310.

The character pixel detecting section 322 detects the number of pixels of the character in the character image of one character cut out by the character cutout section 310 and the ratio of the character pixel to the circumscribed rectangle detected by the circumscribed rectangle detection section 321. (Hereinafter, pixel ratio R) is detected. Then, let the size of the circumscribed rectangle be L,
When A, B, C, and D are threshold values, when A ≦ R ≦ B and L> C, it is determined that only the thinning normalization is performed on the character pixel, A ≦ R ≦ B, and When D ≦ L ≦ C, it is determined that only the enlargement normalization is performed on the character pixel. When A ≦ R ≦ B and L <D, or R <A
Alternatively, when R> B, it is determined that thinning processing is to be performed on the character pixel.

The thinning processing section 323 includes the character pixel detecting section 3
If it is determined at 22 that the thinning processing is to be performed, the thinning processing of the character image of one character cut out by the character cutout unit 110 is performed to obtain the thinned image. next,
The circumscribed rectangle detecting unit 324 detects the circumscribed rectangle of the character from the thinned character image in the thinning processing unit 121, and determines the enlargement or reduction of the image based on the size of the circumscribed rectangle. Note that the method of enlargement / reduction determination is described in the first embodiment.
May be the same as the circumscribed rectangle detection unit 121.

The enlargement normalizing section 325 includes the character pixel detecting section 3
Enlargement normalization processing is performed on the character image whose enlargement is determined by the circumscribed rectangle detection unit 224 or the circumscribed rectangle detection unit 324 to obtain an image whose size is normalized.

The OR normalizing section 326 outputs the circumscribed rectangle detecting section 3
An OR normalization process is performed on the character image for which reduction has been determined in 24 to obtain an image whose size has been normalized.

The thinning-out normalizing section 327 performs thinning-out normalization processing on the character image for which thinning-out determination has been made by the character pixel detecting section 322 to obtain an image whose size has been normalized.

In the thickening processing section 328, the expansion normalizing section 32
5 or the image normalized in size by the OR normalizing unit 326 is subjected to a thickening process so that the width of the line segment of the image becomes a predetermined thickness, and is output to the character recognition unit 130.
Note that the character image that has been determined by the character pixel detection unit 322 to perform only enlargement normalization is output to the character recognition unit 130 without performing the thickening process after the enlargement normalization process performed by the enlargement normalization unit 325. Similarly, the character pixel detection unit 3
The character image determined to perform the thinning process in 22 is output to the character recognition unit 130 without performing the thickening process after the enlargement normalization process in the thinning normalization unit 327.

The character recognizing unit 130 performs character recognition based on the normalized image output from the normalization processing block 320 or the image after the thinning processing, and outputs the character code as a recognition result.

Thus, similar to the first embodiment,
In the normalized image output from the normalization processing block 320, even a portion where the line segment has a large change in thickness is made to have substantially the same line width. Therefore, stable character recognition in the character recognition unit 130 is realized without requiring a plurality of dictionaries corresponding to the line width of characters.

<Embodiment 4> A fourth embodiment of the normalization method of the present invention.
After detecting the circumscribed rectangle of the extracted character image, and performing enlargement normalization, OR normalization, or thinning normalization according to the ratio of the number of pixels of the character image to the circumscribed rectangle,
The normalized image is subjected to a thinning process and the like, and further subjected to a thickening process to make the width of the line segment almost constant. If the character line width does not have a certain thickness and does not adversely affect character recognition, that is, the ratio of the number of pixels of the character to the circumscribed rectangle is smaller than the predetermined value E, and the size of the circumscribed rectangle is Is smaller than the predetermined value F, the thinning process is performed without performing the thickening process. Even when the ratio of the number of pixels of the character to the circumscribed rectangle is between the predetermined values G and H, the processing speed is increased by performing the size expansion normalization processing without performing the thickening processing.

FIG. 4 is a block diagram of an embodiment of a character recognition apparatus to which the above-described fourth normalization method is applied. The character recognition apparatus 40 includes a character cutout unit 110 and a normalization processing block 42.
0, and the character recognition unit 130. The normalization processing block 420 includes a circumscribed rectangle detection unit 421, a character pixel detection unit 422, an enlarged normalization unit 423, and an OR normalization unit 42.
4, a thinning section 425, a thinning section 426, and a thickening section 427.

The character extracting section 110 extracts character images one by one from the image Im read by a character reading device such as a scanner.

The normalization processing block 420 detects a circumscribed rectangle of an image of one character, and performs enlargement normalization, OR normalization, or thinning normalization according to the ratio of the number of pixels of the character image to the circumscribed rectangle. After that, the normalized images are subjected to thinning processing and the like, and further subjected to thickening processing and output to the character recognition unit 130. If the character line width does not have a certain thickness and does not adversely affect character recognition, that is, the ratio of the number of pixels of the character to the circumscribed rectangle is smaller than the predetermined value E, and the size of the circumscribed rectangle is Is the predetermined value F
If it is smaller, the thinning process is performed without performing the thickening process. In addition, even when the ratio of the number of pixels of the character to the circumscribed rectangle is between the predetermined values G and H, the enlargement normalization process is performed without performing the thickening process, and the result is output to the character recognition unit 130.

For this purpose, first, the character extracting unit 110
Circumscribed rectangle detection unit 4 for the character image cut out by
At 21 the circumscribed rectangle of the character is detected, and the enlargement or reduction of the image is determined based on the size of the circumscribed rectangle. The enlargement / reduction determination method may be the same as that of the circumscribed rectangle detection unit 121 of the first embodiment.

The character pixel detecting section 422 detects the number of pixels of the character in the character image of one character cut out by the character cutout section 110 and the ratio of the character pixel to the circumscribed rectangle detected by the circumscribed rectangle detecting section 421. (Hereinafter, pixel ratio R) is detected. Then, let the size of the circumscribed rectangle be L,
When E, F, G, and H are set as threshold values, R ≧ E, or L <F when the circumscribed rectangle detection unit 421 determines that the enlargement is performed.
, It is determined that the thinning processing is performed after performing the expansion normalization on the character pixel. If R <E and the circumscribed rectangle detection unit 421 determines that the character is enlarged, L ≧ F
When it is determined that only the expansion normalization is performed on the character pixel and R <G and the circumscribed rectangle detection unit 421 determines that the character pixel is reduced, after performing OR normalization on the character pixel, It is determined that thinning processing is to be performed, and R> H
If the circumscribed rectangle detection unit 421 determines that the character pixel has been reduced, it is determined that thinning processing is to be performed after thinning normalization is performed on the character pixel. When the reduction is determined by the unit 421, it is determined that only the thinning-out normalization is performed on the character pixel.

The enlargement normalizing section 423 includes the character pixel detecting section 4
An enlargement normalization process is performed on the character image for which enlargement normalization has been determined in 22 to obtain an image whose size has been normalized.

The OR normalizing section 424 is provided for the character pixel detecting section 4
For the character image for which the OR normalization was determined in step 22,
An image whose size is normalized by performing R normalization processing is obtained.

The thinning-out normalizing section 425 applies the character image for which the thinning-out normalization is determined by the character pixel detecting section 422 to
An image whose size is normalized by performing the thinning normalization process is obtained.

The thinning processing section 426 includes the character pixel detecting section 4
If it is determined at 22 that thinning processing is to be performed, thinning processing is performed on the character image whose size has been normalized by the enlargement normalizing section 423, the OR normalizing section 424, or the thinning-out normalizing section. To obtain the thinned image. The thickening processing section 427 performs thickening processing on the thinned image obtained by the thinning processing section 426 so that the width of the line segment of the image becomes a predetermined thickness. Note that the character image determined by the character pixel detection unit 422 to be subjected to only the enlargement normalization is subjected to the enlargement normalization process by the enlargement normalization unit 423 and then to the character recognition unit 130 without performing the thinning process and the thickening process. Output.
Similarly, for a character image determined to be subjected to the thinning process by the character pixel detecting unit 422, the thinning normalizing unit 4
After the enlargement normalization process in step S25, the data is output to the character recognition unit 130 without performing the thinning process and the thickening process.

The character recognition unit 130 performs character recognition based on the normalized image generated by the normalization processing block 420, and outputs the character code as a recognition result.

Thus, similar to the first embodiment,
In the normalized image generated in the normalization processing block 420, a portion where a change in the thickness of the line segment is large has almost the same line width. Therefore, stable character recognition in the character recognition unit 130 is realized without requiring a plurality of dictionaries corresponding to the line width of characters.

[0064]

According to the character image normalizing method and character recognition device of the present invention, the line width of characters in a character image can be made substantially constant. , And stable character recognition by the character recognition unit can be realized, thereby improving the reliability of character recognition.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of a character recognition device to which a first normalization method is applied.

FIG. 2 is a configuration diagram of an embodiment of a character recognition device to which a second normalization method is applied.

FIG. 3 is a configuration diagram of an embodiment of a character recognition device to which a third normalization method is applied.

FIG. 4 is a configuration diagram of an embodiment of a character recognition device to which a fourth normalization method is applied.

FIG. 5 is a diagram illustrating an example of a read image Im.

FIG. 6 is a diagram showing a character image cut out from the image of FIG. 5;

FIG. 7 is a diagram illustrating an example of a character image subjected to a thinning process.

FIG. 8 is a diagram illustrating an example of an enlarged and normalized character image.

FIG. 9 is a diagram illustrating an example of a normalized character image.

[Explanation of symbols]

61 Input Image 71 Character Image 81 Enlarged and Normalized Character Image 91 Character Image of Predetermined Line Width 121, 224, 323, 426 Thinning Processing Unit 122, 221, 422 Bounding Rectangle Detection Unit 123, 222, 325, 423 Enlarge Normalizing sections 124, 223, 326, 424 OR normalizing section (reduced normalizing section) 321 circumscribed rectangle detection section (first circumscribed rectangle detection section) 324 circumscribed rectangle detection section (second circumscribed rectangle detection section) 327, 425 Thinning-out normalization section 328, 427 Thick processing section

Claims (8)

[Claims] 1. A character recognition device for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. And a step of detecting a circumscribed rectangle of the character image after the thinning process and determining whether to enlarge or reduce the character image based on the size of the circumscribed rectangle. A step of performing enlargement normalization processing or reduction normalization processing of the character image; and performing a fattening processing on the character image subjected to the enlargement normalization processing or reduction normalization processing to obtain a character image having a line width of a predetermined thickness. A normalization method for a character image. 2. A character recognition apparatus for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. Detecting and determining whether to enlarge or reduce the character image based on the size of the character image, and performing an enlargement normalization process or a reduction normalization process on the character image based on the determination result. A step of performing a thinning process on the character image that has been subjected to the enlargement normalization process or the reduction normalization process; and performing a thickening process on the character image after the thinning process to obtain a character having a predetermined width. Converting the character image into an image. 3. A character recognition device for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. Detecting, based on the size of the circumscribed rectangle and the ratio of the number of pixels of the character occupying the circumscribed rectangle, to perform only the enlargement normalization or the thinning normalization on the cut-out character image, And performing a normalization process of enlargement, thinning-out normalization, or thinning of the character image based on the result of the determination. In the subsequent stage, a step of detecting a circumscribed circumscribed rectangle of the character image after the thinning process, and determining whether to enlarge or normalize the character image based on the size of the circumscribed rectangle, based on the determination result, Fine Performing an enlargement normalization process or a reduction normalization process on the converted character image; and performing a thickening process on the character image subjected to the enlargement normalization process or the reduction normalization process to obtain a line width having a predetermined thickness. A character image normalizing method. 4. A character recognition device for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. Detecting, from the size of the circumscribed rectangle and the ratio of the number of pixels of the character occupying the circumscribed rectangle, to perform only enlargement normalization or only thinning normalization on the cut-out character image, Determining whether to perform normalization, thinning normalization, or reduction; and performing enlargement normalization processing, thinning normalization processing, or reduction normalization processing of the character image based on the determination result. In addition, the above-mentioned determination result is obtained by expanding normalization, thinning normalization,
Or in the case of reduction, performing a corresponding expansion normalization process, thinning normalization process, or reduction normalization process;
In the subsequent stage, a process of performing a thinning process on the character image that has been subjected to the expansion normalization process, the thinning normalization process, or the reduction normalization process, and performing a thickening process on the thinned character image with a line having a predetermined thickness Forming a character image having a width. 5. A character recognition apparatus for extracting a character image from an input image, performing a normalization process, and then performing a character recognition process on the normalized character image, wherein a thinning process is performed on the extracted character image. Circumscribing rectangle detecting means for detecting a circumscribed rectangle of the character image after the thinning processing, and determining whether to enlarge or normalize the character image based on the size of the circumscribed rectangle; Enlarging normalizing means for performing enlarging and normalizing processing of the character image based on the determination result of the circumscribing rectangle detecting means; and reducing normalization for performing reducing and normalizing processing of the character image based on the judging result of the circumscribing rectangle detecting means. Means for thickening the output of the enlargement normalizing means or the reduction normalizing means to obtain a character image having a line width of a predetermined thickness. Character recognition device according to claim. 6. A character recognition apparatus for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. Circumscribing rectangle detecting means for detecting and determining whether to enlarge or normalize the character image based on its size, and enlarging and normalizing processing of the character image based on the determination result of the circumscribing rectangle detecting means. Expansion normalizing means for performing, reduction normalization means for performing reduction normalization processing of the character image based on the determination result of the circumscribed rectangle detection means, and output from the expansion normalization means or reduction normalization means A thinning processing means for performing thinning processing; and a thickening processing means for performing thickening processing on the output of the thinning processing means to obtain a character image having a line width of a predetermined thickness. Character recognition apparatus, characterized by. 7. A character recognition device for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image, comprising: extracting a circumscribed rectangle of the extracted character image; First circumscribing rectangle detecting means for detecting, from the size of the circumscribing rectangle and the ratio of the number of pixels of the character occupying the circumscribing rectangle, to perform only enlargement normalization or thinning normalization on the cut-out character image Character pixel detecting means for determining whether to perform the thinning or thinning; normalizing processing means for performing expansion normalization processing of the character image based on the determination result of the character pixel detecting means; Reduction normalizing means for performing the reduction normalization processing of the character image based on the result of the determination, and thinning processing means for performing the thinning processing of the character image based on the determination result of the character pixel detection means. And a second circumscribing rectangle for detecting the circumscribed circumscribed rectangle of the character image after the thinning processing and determining whether to enlarge or normalize the character image based on the size of the circumscribed rectangle. A thickening process is performed on a character image enlarged or reduced by the enlargement normalizing unit or the reduced enlargement normalizing unit based on the determination result of the detection unit and the second circumscribed rectangle detection unit, and a predetermined process is performed. A character recognizing device comprising: character normalizing means; and bold processing means for forming a character image having a thickness of a line width. A character recognition device characterized in that: 8. A character recognition device for extracting a character image from an input image and performing a normalization process, and then performing a character recognition process on the normalized character image. Circumscribing rectangle detecting means for detecting, and performing only enlargement normalization or only thinning normalization on the cut-out character image based on the size of the circumscribed rectangle and the ratio of the number of pixels of characters occupying the circumscribed rectangle. Character pixel detecting means for determining whether to perform expansion normalization, thinning normalization, or reduction; and an expansion normalizing means for performing the expansion normalization processing of the character image based on the determination result of the character pixel detecting means. Decimating means, decimating normalizing means for performing decimating normalization processing of the character image based on the determination result of the character pixel detecting means, Reduction normalization means for performing reduction normalization processing of the character image, and further comprising: when the determination result of the character pixel detection means is expansion normalization, thinning normalization, or reduction, Thinning processing means for performing thinning processing on a character image after thinning normalization processing or reduction normalization processing;
A character recognizing device, comprising: character thickening means having thickening processing means for converting the thinned character image into a character image having a line width of a predetermined thickness.