JPH0773269A - Matching method of character recognition result - Google Patents

Abstract

(57) [Summary] [Purpose] Eliminate unnecessary work and simplify work to improve collation efficiency. [Structure] From the original image held in the holding unit 2,
The character image creating unit 4 cuts out individual characters, and the character image reducing / enlarging unit 6 reduces or expands each character so that the size of each character matches the size of the display character font held in the holding unit 5, thereby creating a character. A character image having the same size as the recognition result is displayed on the screen of the CRT 8 at the same time for collation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is a collation system for simultaneously displaying a recognition result of an optical character recognition device (OCR) and an original image of a character on a screen and comparing and collating the recognition result with the original image. Regarding the method.

[0002]

2. Description of the Related Art Conventionally, as a display method of this kind, there is a method as shown in FIG. In FIG. 5, reference numeral a indicates a recognition result, and reference numeral b indicates an original image (original image). That is, FIG.
(B) and (D) are examples where the original image b is larger than the recognition result a, and (C) is an example where the recognition result a is larger than the original image b.

Further, FIGS. 5 (a), 5 (c) and 5 (d) show a case where one line of the recognition result corresponds to one line of the original image, and FIG. 5 (b) shows a plurality of lines (3) of the recognition result. Line) corresponds to one line of the original image. Furthermore, FIG.
(A), (b) and (c) are examples in which the recognition result and the original image are displayed slightly apart, and (d) is an example in which they are displayed close to each other.

As the resolution of image display, the original resolution is often displayed or the image is compressed and displayed at a fixed magnification such as 1/2 times or 1/4 times. It is also common to display a text cursor as shown by symbol c in FIG. 5 (a) and display an original image corresponding to the text cursor in a frame d.

[0005]

However, in the above method, the width of the character image and the width of the character font often do not match, and as a result, the width of the character image is larger than the width of the character font. Figure 5
(A), (b), and (d), and conversely, when the character font width is larger, it becomes as shown in FIG. 5 (c).

Further, the positions of the character and the image to be collated in FIGS. 5A and 5C are deviated, and in such a case,
There is a problem in that the collation work must be performed while alternately confirming two different places on the screen, and the work is troublesome.
Further, when the character to be collated and the image are displayed close to each other at the position of the text cursor as shown in FIG. 5D, it is necessary to move the text cursor on the screen along the character to be collated. If the image width is larger than the character font width, only a part of the recognition result image is displayed. From this point as well, the work of moving the text cursor along the collation target character is required.

That is, in the conventional method, there is a problem that the work for coping with the position shift or the work for the position alignment is required in addition to the original collation, which increases the load. Therefore, an object of the present invention is to enable collation work in a short time without requiring extra work and improve collation efficiency.

[0008]

In order to solve such a problem, according to the present invention, the result of recognizing a document is held, while the original image of the document itself and the position and size information of each character in the original image are stored. After holding and, individual characters are cut out from the original image, and each cut out character image is reduced or enlarged to match the character font size of the recognition result acquired in advance. The feature is that the characters are displayed close to each other and displayed on the screen at the same time.

[0009]

[Operation] Matching is performed by matching the height and width of each character of the original image with the height and width of the previously acquired character font and displaying the obtained image close to the character font of the recognition result. Streamline work and improve work efficiency.

[0010]

1 is a block diagram for explaining an embodiment of the present invention. In the figure, 1 is a character recognition result holding unit, 2 is an original image holding unit, 3 is a character position / size information holding unit, 4 is a character image creating unit, 5 is character font information holding unit, and 6 is a character image reduction unit. An enlargement unit 7, a recognition result / image display unit, and a display device (CRT) 8 are shown.

That is, the character recognition result holding unit 1 holds the recognition result obtained by reading the document, and the character position / size information holding unit 3 holds the position, size, etc. of each character obtained in the recognition process on the original image. Of the input image information, and the original image holding unit 2 holds the input image information by binarizing it.
Further, the character font information holding unit 5 stores information regarding characters to be displayed on the CRT 8.

Therefore, in the character image creating section 4,
Each character is cut out from the original image given by the original image holding unit 2, and the size (character size) of the cut out character image is cut by the character image reduction / enlargement unit 6 based on the information given by the character font information holding unit 5. Is reduced to fit the size of the display character font. Usually, it is often reduced, but if the character image is smaller, it is enlarged.

The recognition result / image display unit 7 receives the information of the character recognition result and the character image information from the character image reduction / enlargement unit 6 and performs predetermined control to display them on the CRT 8 to The recognition result and the character image corresponding to the recognition result are displayed on the CRT 8 in the same size and close to each other.

Now, for example, assume that the original image is that shown in FIG. 2, and one of the CH (sub) is cut out by the character image creating unit 4. Assuming that the character font has a height H, a width W, and a start position of a character portion as shown in FIG. 3, the size (height H1, width W1) of the character image "A" is the height of the character font. Compress to fit H and width W. Here, as shown in FIGS. 2 and 3, the character font is shown larger, but in reality, the character font is often smaller, and therefore the character image is compressed in most cases.

The image thus compressed is displayed as shown at b at a position shifted from the character start position S1 by "S" shown in FIG. 3, as shown in FIG. a indicates the recognition result.
Thus, by displaying the reduced and enlarged original image close to the character font of the recognition result, the collation work is rationalized and the work efficiency is improved. Although an example of horizontal writing has been described above, it goes without saying that the present invention can be similarly applied to the case of vertical writing.

[0016]

According to the present invention, since the character recognition result and the original image of each character can be displayed in substantially the same size and in a state where they are close to each other in a one-to-one manner, the collation work can be performed vertically and horizontally. Or you can just compare the columns,
As a result, the work for collation is simplified, and the work time can be significantly shortened.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating an example of an original image of a document.

FIG. 3 is an explanatory diagram illustrating an example of a character font.

FIG. 4 is an explanatory diagram for explaining an example of a matching screen according to the present invention.

FIG. 5 is an explanatory diagram for explaining a conventional method.

[Explanation of symbols]

1 ... Character recognition result holding unit, 2 ... Original image holding unit, 3 ...
Character position / size information storage unit, 4 ... Character image creation unit, 5 ... Character font information storage unit, 6 ... Character image reduction / enlargement unit, 7 ... Recognition result / image display unit, 8 ... Display device (CRT), a … Character recognition result, b… Original image.

Claims (1)

[Claims] 1. The document recognition result is held, while the original image itself of the document and the position and size information of each character in the original image are held, and then individual characters are cut out from the original image, and this cutting out is performed. Characters that are displayed on the screen at the same time by reducing or enlarging each character image so that it matches the character font size of the recognition result acquired in advance, and making it closer to the character of the recognition result. How to match the recognition results.