JPH05342409A - Character recognizing device - Google Patents

Abstract

PURPOSE:To correctly recognize characters protruded out of character entry frames. CONSTITUTION:This device is provided with a character segmenting means to select an area larger than the entry frame from slip images as the segmenting area of characters in the case of segmenting an area equivalent to one character by using the size information of the description frame stored in an entry frame size information part 114 at a character segmenting part 112 under the control of a control part 115 from the images of a reading object slip stored in a slip image memory 111 of a character segmenting means 11 by a photoelectric converting means 10 of a character pattern input means 1. Thus, even the character protruded out of the entry frame can be segmented without being partially missed, this is recognized by a character recognizing means 2, and the machine code of the relevant character is outputted.

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 for reading a character pattern written on a form and recognizing a character from the read character pattern.

[0002]

2. Description of the Related Art In a conventional character recognition device, an entry frame is set on a form in order to specify a character entry position. At the time of writing characters in this entry frame, there is no longer any restriction on the typeface for the writer as shown in the sample characters, but the entry frame has the effect of suppressing fluctuations in the size of the letters. It is well known that it contributes to the realization of stable reading performance for free-form fonts. As a reference, Michio Yasuda: “How is a character reader configured?”, Nikkei Electronics 197
5.1.27, and Shinichiro Hashimoto: “Introduction to Character Recognition”, Telecommunications Association. The entry frame is one of the entry restrictions in the character recognition device, and requires the writer to enter characters that do not extend beyond the entry frame. Therefore, it is premised that the characters are entered without protruding into the entry frame, and each character recognizes the image of the portion corresponding to the inside of the entry frame by cutting out from the form.

[0003]

In the above-described conventional character recognition device, characters often protrude from the entry frame in actual operation, and misreading occurs in the conventional method in which only the inside of the entry frame is cut out and recognized. There is a drawback that there are cases where

This example will be described below with reference to FIG.
Since the conventional character recognition device cuts out the inside of the entry frame, the character patterns that are entered outside the entry frame as shown in the upper part of FIG. 5 have the cutting results shown in the middle part of FIG. 5, respectively. The recognition result when these are sent to the character recognition means is, for example, as shown in the lower part of FIG. 5, and correct reading cannot be performed. This is the result of ignoring the outside of the entry frame, as you can see from the cutting result in the middle of Fig. 5,
This is because the pattern of the portion corresponding to the inside of the remaining entry frame resembles another character.

An object of the present invention is to provide a character recognizing device which can correctly read a character pattern which is out of the entry frame as described above.

[0006]

A character recognition device of the present invention is a character pattern input means for converting a character pattern written on a form into an electric signal and cutting out characters using the form pattern converted into the electric signal. In the character recognition device, the character pattern input means converts the character pattern into an electric signal, and prints on the form to indicate the writing position of the character using the form pattern converted into the electric signal. When cutting out the characters entered in the entered entry frame,
Characters can be cut out by using an area larger than the entry frame on the form as a cutout area for each character.

[0007]

The operation of the present invention will be described with reference to the example shown in FIG.

Regarding the characters that are written out of the character entry frame, if a conventional cutout method causes a reading error, only the size of the character is the main cause of the reading error. The problem can be solved by cutting out a part larger than the frame as the cutout target area. In contrast to the form entry example in the upper part of FIG. 4, for example, an area including the peripheral portion of the entry frame can be targeted for character extraction as indicated by a dashed line. By setting such a cutout area, a character image without partial omission as shown in the middle part of FIG. 4 can be obtained. As a result, as shown in the lower part of FIG. 4, correct reading is performed in the character recognizing means without newly adding a reading method for such a partially missing character pattern.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 2 is a block diagram of a character recognition device showing an embodiment of the present invention. The character recognition device of FIG. 2 includes a character pattern input means 1 and a character recognition device 2.

The character pattern input means 1 is means for taking in an image written on a form. Further, the character pattern input means 1 includes a photoelectric conversion means 10 for converting a character image into an electric signal, and a character cutting means 11 for cutting out a portion to be read for one character from the image Y1 converted into the electric signal. Composed. Of these, a normal image scanner may be used as the photoelectric conversion means 10.

The character cut-out means 11 has a form image memory 111, a character cut-out 112, as shown in FIG.
It includes a cutout data buffer 113, an entry frame size information storage unit 114, and a control unit 115.

The form image memory 111 is a memory for storing electric signals corresponding to the form image converted by the photoelectric conversion means 10 shown in FIG. 2, and has a storage capacity capable of storing an image of one form. .. The character cutout unit 112 is a portion that cuts out a read target area for one character from the image stored in the form image memory 111, and can cut out an area larger than the character frame filled in the form. It is possible.

FIG. 3 shows a flow of operation in which the character cutout unit 112 cuts out a read target area for one character. In FIG. 3, it is determined whether or not the first character line entered is out of the entry box (steps 31, 3).
2) If there is no protrusion, only the inside of the character frame is set as the target region for clipping (step 33). If the character line is out of the entry frame in step 32, the circumscribing frame of the entered character is determined (step 34), and the circumscribing frame is set as the cutout area (step 35).

In step 32, it is determined whether or not the character line is out of the character frame by using the form image memory 111 on the control unit 1115 by checking whether or not the written character line overlaps with a portion corresponding to the character frame. Just make a decision. For the circumscribing frame for the entered characters, for example, the maximum value in the horizontal direction, the minimum value, the maximum value in the vertical direction, and the minimum value in the vertical direction where the character lines exist are detected by scanning in the vertical direction and the horizontal direction with the center of the input frame as a reference. To obtain. At this time, without judging whether the character frame is protruding or not, the circumscribing frame of the character pattern is detected in advance for all the input frames,
You may cut out. In addition, the same effect can be obtained by setting a large cutout area in advance so that it does not come into contact with an adjacent character frame.

The entry box size information storage section 114 is a section for storing the size of the character box printed on the form,
The area for one character is referred to when setting the cutout area.
The cutout data buffer 113 also stores the character image cut out by the character cutout unit 112 as cutout data. Then, the control unit 115 causes the above-mentioned 11
It controls each device from 1 to 114.

The character recognizing means 2 is a part for outputting a machine code Y3 of a corresponding character type from the image Y2 for one character cut out by the character pattern inputting means 1, and the method is appropriately selected according to the object to be read. do it.

Further, in the present embodiment, one character is written in one character frame, but it is also applicable to an entry field provided when collectively writing a character string such as an address or a telephone number. Is possible.

As described above, by cutting out a character from an area larger than the character frame, it is possible to correctly read the character that has been written out of the entry frame. Further, in the character recognition means of the character recognition device having the present invention, it is not necessary to design in consideration of the partial omission of the character pattern due to the above-described factors described in FIG. The load and the scale of the device are reduced.

[0019]

As described above, according to the character recognition device of the present invention, by cutting out a character from an area larger than the character frame, it is possible to correctly read the character that has been written out of the entry frame. This has the effect of improving performance.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a character cutting means of a character recognition device of the present invention.

FIG. 2 is a block diagram showing an embodiment of a character recognition device of the present invention including the character cutting means shown in FIG.

FIG. 3 is a flowchart showing an example of an operation of the character cutting means shown in FIG.

FIG. 4 is an explanatory diagram for explaining the operation of the character recognition device of the present invention.

FIG. 5 is an explanatory diagram for explaining the operation of the conventional character recognition device.

[Explanation of symbols]

 1 Character pattern input means 2 Character recognition means 10 Photoelectric conversion means 11 Character cutout means 111 Form image memory 112 Character cutout section 113 Cutout data buffer 114 Entry frame size information storage section 115 Control section

Claims (1)

[Claims] 1. A character recognition device comprising character pattern input means for converting a character pattern written on a form into an electric signal and cutting out characters using the form pattern converted into the electric signal. The means converts the character pattern into an electric signal, and uses the form pattern converted into the electric signal, and cuts out the character written in the entry frame printed on the form to indicate the writing position of the character. At this time, the character recognition device is characterized in that it is possible to perform character extraction by using an area larger than the entry frame on the form as an extraction area for each character.