JPS6237786A - Character segmenting device - Google Patents

Abstract

PURPOSE:To reduce the processing time and to improve the processing efficiency with a character segmenting device, by changing the processing contents according to the state of arrangement of characters at a place close to the boundary between characters and therefore discriminating the arrangement of characters at the place close to the character boundary. CONSTITUTION:A picture reading device 11 of a character segmenting device reads the picture information on the surface of a document and supplies them to the segmenting means 12-15 respectively. These means 12-15 segment individual character patterns on the surface of the document and supply the character segmentation result signals to a character segmentation control means 16. This means 16 discriminates the arrangement of characters at a place near the boundary between the character patterns to be segmented next according to the picture data. Based on the result of said discrimination, the control signal is produced for selection of one of means 12-15. The control signal is supplied to the means 12-15 respectively. Thus these means 12-15 are actuated to deliver the output character pictures. Thus the processing efficiency is improved with a character segmenting device.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a character segmentation device for an optical character reading device.

BACKGROUND OF THE INVENTION In recent years, optical character reading devices that use an offline character recognition method as a character input method have been put into practical use in order to store, process, search, etc. in document information computers. In optical character reading devices that read characters, a character segmentation device that detects and segments each character prior to character identification is essential.

An example of the conventional character cutting device mentioned above will be described below with reference to the drawings. FIG. 4 shows a block diagram of a conventional character cutting device. In FIG. 4, 41 is an image reading device, 42 is a character line extraction means, and 43 is a character line integration means.

The operation of the character cutting device configured as described above will be explained below.

First, the image reading device 41 reads information on the paper of the document as image data, and transmits the image data to the character line extraction means 42. Character line extraction means 42 extracts character lines near character boundaries and transmits character line information to character line integration means 43. The character line integration means 43 cuts out characters and outputs a character image by determining whether each character line belongs to the left or right side of the character boundary in response to the character line information (for example, the IEICE technical research report pattern recognition) and learning PRL8
0-5).

Problems to be Solved by the Invention However, the above configuration has a problem in that the same character line extraction process is performed for all characters, which requires a lot of effort.

Means for Solving the Problems In view of the above-mentioned problems, the present invention has a structure that reduces processing effort by changing the processing content in accordance with the arrangement of characters near character boundaries. .

Function: With the above-described configuration, the present invention can determine the character arrangement near the character boundary and apply processing corresponding to the character arrangement, thereby reducing processing effort.

EXAMPLE Hereinafter, a character cutting device according to an example of the present invention will be described with reference to the drawings. FIG. 1 shows a block diagram of a character segmentation device according to a first embodiment of the present invention. In FIG. 1, 1. is an image reading device; 12.
13, 14 and 15 are cutting means, and 16 is a character cutting control means.

The operation of the character cutting device configured as described above will be explained below with reference to FIGS. 1 and 2.

First, FIG. 2 shows a part of a character string, where 21 is a character string, 22 is a character boundary, and 23 is a reference range.

The image reading device 1) inputs a character string 21 and converts the corresponding image data into a character cutting control means 16 and a cutting means 1.
2.13.Transmit on 14 and 15.

The character cutting control means 16 first transmits a control signal indicating that one character should be cut out to the cutting means 12, and the cutting means 12 performs cutting in response to the signal. If it is determined that the cutting means cannot cut out the character, A character cutting result signal indicating that the character cannot be cut out is transmitted to the character cutting control means 16, and the character cutting control means 16 then activates the cutting means 13. The cutting means 13 performs cutting in response to the signal, and when it is determined that the cutting means cannot cut out the character, the character cutting means 13 sends a character cutting result signal indicating that the character cannot be cut out to the character cutting control means 1.
6, and the character cutting control means 16 then activates the cutting means 14. The cutting means 14 performs cutting in response to the signal, and when it is determined that the cutting means cannot cut out the character, transmits a character cutting result signal indicating that the character cannot be cut out to the character cutting control means 16,
The character cutout control means 16 then activates the cutout means 15. The cutout means 15 performs cutout in response to the signal, outputs a character image, and transmits a character cutout result signal indicating that one character has been cut out to the character cutout control means 16.

When each of the cutout means 12, 13 and 14 determines that a character of the character string 21 can be cut out, the cutout means 12, 13 and 14 cut out one character and output a character image, and notify the character cutout control means 16 that one character has been cut out. The character segmentation result signal indicated is transmitted. Character cutting control means 16
In response to the signal, the character string 21 transmits a control signal to the extraction means 12 indicating that the next character on the character string 21 should be extracted.

As described above, according to this embodiment, by providing a plurality of cutting means and a character cutting control means for controlling the cutting means in response to a character cutting result signal transmitted by the cutting means, processing corresponding to the character arrangement is performed. can be added, which reduces processing time.

A second embodiment of the present invention will be described below with reference to the drawings.

FIG. 3 is a block diagram of a character cutting device showing a second embodiment of the present invention. In the figure, 31 is an image reading device, 32, 33, 34 and 35 are cutting means, respectively.
36 is a control means. The difference from the configuration shown in FIG. 1 is that the cutout means 32, 33, 34, and 35 are provided so as not to transmit the character cutout result means.

The operation of the character cutting device configured as described above will be described below.

The image reading device 31 inputs the character string 21 and sends the corresponding image data to the control means 36 and the cutting means 32, 33, 34.
and 35.

The control means 36 determines the arrangement of characters in the reference range 23 of the character string 21, and if there is no character line on the character boundary 22, the control means 36 sends a control signal to the extraction means 32 to activate the extraction means 32.
to be transmitted. In response to the control signal, the cutting means 32 cuts out characters along scanning lines perpendicular to the character string at character boundaries 22 and outputs a character image.

The control means 3G scans the character string 21 in the reference range 23 when there is a character line on the character boundary 22, and sends a control signal to activate the cutting means 33 when there is a scanning line without a character line in the reference range 23. It is transmitted to the extraction means 33. In response to the control signal, the cutting means 33 cuts out a character at the scanning line closest to the character boundary 22 among the scanning lines without a character line in the reference range 23, and outputs a character image.

The control means 36 transmits a control signal for activating the cutting means 34 to the cutting means 34 when character lines are present on all the scanning lines of the reference range 23 . The cutting means 34 tracks connected black dots existing in the reference range 23 in response to the control signal, and tests the relationship between each set of connected black dots and the left and right ends of the reference range 23. If there is no connected sunspot set that touches both the left and right ends of the reference range 23, each sunspot set is in the reference range 2.
Corresponding to the case where it does not touch either the left end or right end of 3,
It is determined that each black dot set belongs to the character on the left side of the character boundary 22, the character on the right side, or none. The cutout means 34 cuts out one character based on the above judgment and outputs a character image.

The control means 36 activates the extraction means 35 when there is a connected set of black points touching both the left and right ends of the reference range 23. The cutting means 35 cuts out one character along the scanning line on the character boundary 22 and outputs a character image.

The control means 36 control the cutting means 32, 33, . . . as described above.
After activating either 34 or 35, the next character extraction process is started.

Effects of the Invention As described above, by determining the character arrangement of a character string in the control means and activating the corresponding extraction means, the configuration of the extraction means can be simplified and processing efficiency can be improved. can.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a character segmentation device in a first embodiment of the present invention, FIG. 2 is a configuration diagram of a character string that may appear in a document, and FIG. 3 is a character segmentation device in a second embodiment of the present invention. Block Diagram of Device FIG. 4 is a block diagram of a conventional character cutting device. 1)... Image reading device, 12.13.14 and 15... Cutting means, 16... Control means, 21... Character string, 22. ... Character boundary, 23 ... Reference range, 31 ... Image reading device, 32, 33, 34 and 35 ... Cutting means, 36 ... ...control means, 41...
. . . Image reading device, 42 . . . Character line extraction means,
43...Character line integration means. Name of agent: Patent attorney Toshio Nakao and one other person Figure 1 (Character t) Figure 6 and Figure 2 ■ ??Character Boundary Figure 3 Figure 4

Claims (1)

[Scope of Claims] 1) an image reading device that reads and outputs information on a document page as image data; and an image reading device connected to the image reading device;
a plurality of cutting means for cutting out individual character patterns on the document surface in response to the image data supplied from the image reading device; determines the character sequence near the boundary of the character pattern to be cut out next based on the image data of A character cutting device comprising a character cutting control means for controlling the operation of the cutting means by supplying a character to the cutting means. (2) The cutting means scans the character string in a direction perpendicular to the character string at a constant pitch, and if there is no character line on the scanning line, cuts out the character along the scanning line and outputs the character image; A character cutting result signal indicating that one character cutting has been completed is transmitted to the character cutting control means, and if a character line exists on the scanning line, the character cutting result signal indicating that the character cannot be cut out by the cutting means is transmitted to the character cutting control means. A character cutting device according to claim (1), characterized in that the character cutting device is configured to transmit the information to a character cutting control means. (3) The cutting means predicts a character boundary based on a constant pitch, scans the character string in a direction perpendicular to the character string with a constant width centered on the character boundary, and places a character line on the scanning line among the scanning lines. If there is a scanning line in which the character line does not exist, the character is cut out at the scanning line closest to the character boundary among the scanning lines in which the character line does not exist, and the character image is output, and the character cutting control means finishes cutting out one character. A character cutting result signal indicating that the character cannot be cut out by the cutting means is transmitted to the character cutting control means if character lines are present in all of the scanning lines. A character cutting device according to claim (1), characterized in that: (4) The cutting means predicts character boundaries based on a constant pitch, and traces character lines in an area of a constant width centered on the character boundary, so that each character line is placed on either the left or right side of the character boundary. If the attribution of all character lines can be determined, the character is extracted based on the attribution of the character line and a character image is output, and the character extraction control means finishes the extraction of one character. If there is a character line whose belonging cannot be determined, the character extraction result signal indicating that the character cannot be extracted by the extraction means is transmitted to the character extraction control means. A character cutting device according to claim (1), characterized in that: (5) The cutout means cuts out characters along scanning lines perpendicular to the character string based on a constant pitch, outputs a character image, and sends a character cutout result signal to the character cutout control means to notify that one character has been cut out. A character cutting device according to claim 1, wherein the character cutting device is configured to transmit.