JPH01111285A - Character recognition result confirmation/correction processing system for graphic recognizing processing - Google Patents

Abstract

PURPOSE:To simplify an operation for confirmation/correction by making a recognized character and a character segmented image from an original image into one pair, displaying plural pairs of them at one time, and confirming and correcting a character recognition result. CONSTITUTION:Plural pairs of recognized character display areas 25 and character detection image display areas 26 are provided on a display screen 24. Further, a recognized character display processing part 11 makes a recognized character column inputted from a graphic input device 16 correspond to the segmented image, and displays plural pairs. A recognized character correction processing part 12 operates a cursor 28 on plural areas 25 by the input from a key board 20, a tablet 21, etc., and performs a processing to correct the erroneously recognized character and unrecognizable character to the correct character. When it is difficult to judge to what part of the graphic the segmented image of the recognition object displayed in an area 26 corresponds, since the peripheral image can be displayed on a peripheral image display area 27, the confirmation/correction is effectively attained.

Description

[Detailed Description of the Invention] [Summary] Character recognition result confirmation/correction processing method in drawing recognition processing that allows single character recognition results in a drawing input system to be easily and easily confirmed and corrected.Man-machine related to confirmation/correction of character recognition results. The purpose is to improve the interface and reduce the effort and time required for confirmation and correction.

A plurality of sets of recognized character display areas and character extraction image display areas are provided on a display screen, and the recognized character display areas and the character extraction image display areas are provided.

Recognized character display processing means that displays a recognized character result and an image to be recognized in correspondence for each recognized character string, and a recognized character that updates the recognized character by modifying the screen in the recognized character display area. The present invention is configured to include a correction processing means and a surrounding image display processing means for displaying a surrounding image including a recognized character to be corrected in another area within the same screen.

[Industrial application field]

The present invention uses the character recognition results in the 9 drawing input system,
The present invention relates to a character recognition result confirmation/correction processing method in drawing recognition processing that enables easy-to-understand and easy confirmation/correction.

A drawing input device is used that extracts and recognizes nine characters from an input drawing read by a scanner, and outputs image data, recognized character data, and the like. For characters recognized by such devices.

Especially with handwritten drawings, recognition errors may occur.
It is necessary to confirm and correct it. A man-machine interface that can efficiently perform this confirmation and correction is desired.

[Conventional technology]

FIG. 7 shows an example of the conventional method.

In FIG. 7, 24 represents a display screen, 50 represents an image display field on the display screen, and 51 represents a character input field.

Conventionally, when checking and correcting character recognition results in drawing recognition processing, 1. For example, as shown in FIG. 7, the display screen 24
An image display field 50 for displaying an image based on the read dot pattern, and a character input field 51 for displaying characters in units of recognized character strings and, if there is an error, for inputting on the screen in that area, are provided at the top. The entire original image and the individual character recognition results were displayed, and each recognized character was checked and corrected individually.

[Problem that the invention seeks to solve]

According to the above-mentioned conventional method, the original image and the recognized characters are displayed and the two are checked and corrected individually, which makes operations such as moving and enlarging the original image within the display complicated.

Since confirmation and correction must be performed one by one for each recognized character string, there is a problem in that the confirmation and correction takes a long time.

The present invention aims to solve the above-mentioned problems and improve the man-machine interface for checking and correcting three-character recognition results, thereby reducing the labor and time required for checking and correcting results.

[Means for solving problems]

FIG. 1 is a block diagram of the principle of the present invention.

In FIG. 1, 10 is a processing device consisting of a CPU, memory, etc., and 11 is a recognized character display processing section.

12 is a recognized character correction processing unit; 13 is a peripheral image display processing unit; 14 is a recognized character table containing recognized character strings and their extraction coordinate information; and 15 is an image expansion memory in which the read original image is expanded.

16 is a drawing input device, 17 is a scanner that reads the drawings,
Reference numeral 18 indicates a figure recognition device that recognizes figures based on read images, 19 indicates an input device, 20 indicates a keyboard, 21 indicates a tablet, and 22 indicates a primary storage device in which image data, recognized character data, vector data, etc. input from the drawing input device 16 are stored. 23 is a display, 24 is a display screen, 25 is a recognized character display area where the recognized character result is displayed, 26 is a character extraction image display area where the image to be recognized is displayed, and 27 is the current focus. A peripheral image display area 28 displays a peripheral image of the recognized character result. Reference numeral 28 represents a cursor indicating the screen position.

Image data, recognized character data, vector data, etc. of drawings input from the drawing input device 16.

The information is stored in the primary recognition file 22, which is an external storage medium managed by the processing device 10. This recognition figure.

In particular, when checking and correcting the character recognition results via the display 23, the image data is developed in the image development memory 15, and the recognized character data is developed on the memory as the recognized character table 14.

The recognized character display processing unit 11 displays a plurality of sets of recognized character results and images to be recognized in a recognized character display area 25 and a character extraction image display area 26 on the display screen 24 in units of recognized character strings. This process performs a process of matching and displaying.

The recognized character modification processing section 12 performs a process of updating the recognized character information in the recognized character table 14 by modifying the screen in the recognized character display area 25.

The peripheral image display processing section 13 performs a process of displaying a peripheral image including the recognized character to be corrected in a peripheral image display area 27 prepared in another area on the same screen. That is, the recognized character string displayed in the recognized character display area 25 where the cursor 28 is present is 1.
The partial image on the image expansion memory 15 is displayed so as to be located approximately in the center of the peripheral image display area 27.

[Effect]

In the present invention, a plurality of sets of recognized character display areas 25 and nine character extraction image display areas 26 are provided in the display screen 24, and the recognized character strings input from the two drawing input devices 16 by the recognized character display processing section 11 are provided. and its cut-out images are associated with each other and displayed in a plurality of sets within the size of one screen.

In contrast, the recognized character correction processing unit 12
0. By inputting from the tablet 21 or a mouse (not shown), the cursor 28 is moved over the plurality of recognized character display areas 25, and misrecognized characters and unrecognized characters replaced with, for example, "?" are corrected to correct characters. Perform processing.

Therefore, a plurality of recognized character strings can be confirmed and corrected on one screen.

When the correction of misrecognized characters etc. for one screen is completed,
Alternatively, if the recognized character is correct, the recognized character display processing section 11 displays a plurality of sets of secondary recognized character strings and their cut-out images in response to instructions from the input device 19.

If it is difficult to understand which part of one drawing the cutout image of the recognition target displayed in the character extraction image display area 26 corresponds to, the second peripheral image display processing unit 13 displays the surrounding area on the first peripheral image display area 27. Since the image can be displayed, confirmation and correction can be easily and efficiently performed.

[Example], Figure 2 is an example of recognized character data in an embodiment of the present invention, Figure 3 is an example of a processing configuration according to an embodiment of the present invention, and Figure 4 is confirmation in an embodiment of the present invention. An example of a correction screen, FIG. 5 shows an example of confirmation and correction according to an embodiment of the present invention, and FIG. 6 shows an example of horizontal character string confirmation and correction processing according to an embodiment of the present invention.

First, an example of data handled in this embodiment will be explained with reference to FIG.

As shown in FIG. 2(a), the single drawing input device 16 reads the second drawing 30, analyzes the image data, cuts out an image that is considered to be a character string from the original drawing, and performs single character recognition. Furthermore, image data is converted into vector data consisting of a collection of coordinate data such as line segments.

This type of drawing input device 16 has been put into practical use.

Since this is well known, I will limit the explanation to this level.

The processing device 10 receives data from two drawing input devices 16, such as image data compressed by MH encoding, recognized character data,
It also receives vector data consisting of coordinate information such as start and end points, line type and other attribute information. Particularly, the recognized character data related to the present invention includes the coordinate information of the cut-out image, the character code of the recognition result, etc.

The information on the cut-out image to be recognized is 2 For example, as shown in FIG. This is the coordinate information of point P at the upper left corner of , and its diagonal point Q.

As for the recognized character data, as shown in Figure 2 (C),
xy coordinate (xi, yl) information of point P.

The xy coordinates (x 2. y 2) information of point Q, horizontal/vertical flag information indicating whether one character is an upright character or a horizontally turned character, information on the number of characters in the recognized character string, and recognition as the recognition result. Contains string code information, etc.

These are organized into tables and managed. The recognized character data shown in FIG. 2 (c) is.

It is repeated for the number of recognized character strings.

FIG. 3 is an example of a schematic configuration of processing in this embodiment.

In the main screen process 41, a menu of various processes related to confirmation and correction is displayed and the process corresponding to the selected menu item is activated. For example, the horizontal character string confirmation/correction process 42 is activated by inputting the function key PFI. Further, by inputting the function key PF2, a vertical character string confirmation/correction process 43 is activated. With the function key PF3, a confirmation/correction process 44 using the entire original image is activated. In the confirmation/correction processing 44 using the original image portion, confirmation/correction processing similar to the conventional method is mainly executed.

In the horizontal character string confirmation/correction process 42, for example, a display screen 24 is used as shown in FIG. 4(A). A total of eight sets of recognized character display areas 25 and character extraction image display areas 26 are provided on the left and right, top and bottom, and five peripheral image display areas 27 are provided at about 1/3 of the bottom of the nine screens.

In the vertical character string confirmation/correction process 43, for example, the display screen 24 is used as shown in FIG. 4(b). A total of seven groups of recognized character display areas 25 and character extraction image display areas 26 are provided above and below.

Nine peripheral image display areas 27 are provided vertically on the right side of the screen.

Processing by horizontal character string confirmation/correction processing 42.

What is the process performed by the vertical character string confirmation/correction process 43?

Since the contents are the same, an example of the horizontal character string confirmation/correction process 42 will be described below.

In this embodiment, function keys are used in the following manner in this confirmation/correction process.

PFI: Display of peripheral image PF2: Enlarge peripheral image PF3: Correct string extraction PF5: Proceed to the next confirmation correction screen PF6: Exit the confirmation/correction screen PF7. PF8. PFIO, PFII :Scroll surrounding images PF12: Switch between original size/2x enlargement of image characters PF17: Return one confirmation/correction screen.

In the horizontal character string confirmation/correction process 42, first, based on the recognized character table 14 shown in FIG.
As shown in the figure, the results of the first 8 klI of recognized characters and their cut-out images are displayed.

Here, in the actual string rABcDJ.

rBJ was mistakenly recognized as "8", and rDJ was also unable to recognize the unrecognized character.

Here, when the display of the peripheral image 3 is instructed by the function key PFI, for example, and the recognized character display area 25 is hit by a tablet or the like.

Figure 5 (b) 1 peripheral image display area 27 as shown
A surrounding image including the one character string cutout image rABcDJ is displayed. Especially regarding the character string cutout image part in one peripheral image.

It is displayed with red shading so that it can be visually identified.

When the function key PF2 is pressed, this peripheral image is enlarged twice and displayed.

For example, as shown in FIG. 5(c), the recognized character display area 2
5, place the cursor at the "8" position and enter "1B", and place the cursor on "?" and enter rDJ to change the recognized character string to the correct character string rABcD.
It can be modified to J.

For example, in reality, a series of character strings rP80-135J is changed to rP80-J due to the spacing between two characters.

There are cases where it is recognized as two character strings: rl 35J.

In this case, by pressing the function key PF3, it is possible to instruct correction of the 9-character cutout portion. For example, Figure 5 (d)
As shown in , after pressing function key PF3,
Diagonal point P in the rectangular area of the image part you want to change,
Hint Q.

Change the crop area to a new rectangular area.

Drawing input devices often incorrectly recognize characters such as r-J, r.J, r, J, etc. in blank areas due to dirt or noise in the drawing, for example.

In this embodiment, these recognition results are also displayed in the recognized character display area 25, and by overwriting the "blank" character,
It allows you to delete those characters. When a character is deleted, "" is displayed in the recognized character display area 25 for confirmation.
DELETE, J is displayed.

After checking and correcting the recognized character string for one screen.

By pressing the function key PF5, eight secondary sets of recognized character strings are displayed, and confirmation and correction can be continued.

The internal processing by the horizontal character string confirmation/correction process 42 is as shown in FIG. 6, for example. Here, corresponding to the recognized character display area 25, the recognized character work W1 in which eight pieces of recognized character string code information are stored, and eight character extraction coordinates corresponding to the nine character extraction image display area 26. The work area of the recognized character cutting work W2 in which information is stored is used. What is the data structure stored in these works?

It can be considered that the data structure is such that it has recognition character string information and image coordinate information similar to the data structure shown in FIG. 2(c).

In the following explanation, ■ to [phase] are the processes shown in FIG.
Corresponds to ~0. Furthermore, at the time of this processing.

Recognized character table 14 shown in FIG. As the image development memory 15, information such as images, recognized character strings, and the extraction coordinates of each character string is developed in a virtual memory in an image coordinate system.

■ Select eight recognized character strings from the recognized character table 14 shown in FIG. 1, write them on the recognized character work W1, and display them on the character screen of the display 23.

Create and display a recognition character field.

■ Also, set the coordinate information regarding the cut-out images of the eight recognized character strings displayed in process ■ in the recognized character cut-out work W2, create a hard command to display that image, and create a one-character extraction image display area 26. to be displayed.

■ Wait for input from a keyboard, tablet, etc.

■PF5. PF6. ' Perform the following processing depending on the input type such as PFI 7.

(2) When the function key PF17 or PF5 is input, the recognized character work W1 and the recognized character cutting work W2 are referred to, and the corrected recognized character and its cutting coordinates are reflected in the recognized character table 14.

(2) When the function key PF17 is input, the pointer to the recognized character table 14 is moved back eight positions and the screen is changed to the previous screen. Also, if PF5 is input,
The pointer to the recognized character table 14 is moved forward by eight positions to change to the primary screen.

■ Clear the screen, return control to process ■, and repeat the process for the previous or next screen.

■ Function key PF5. PF6. If a value other than PFI7 is input, the process corresponding to that input is executed. In particular, when a character is input in the recognized character display area 25, the contents of the recognized character work W1 are updated. When the two-character extraction is corrected using the function key PF3, the contents of the recognized character work W2 are updated. After executing the corresponding process, control is returned to process (2) and waits for the secondary input.

■ When function key PF6 is input,
The recognized character work W1 and the recognized character cutout work W2 are referred to, and the recognized characters that have been modified so far and their cutout coordinates are reflected in the recognized character table 14.

[Phase] Clear the screen and perform main screen processing 41 shown in Figure 3.
Return control to.

The detailed explanation of the present invention has been given above regarding the confirmation and correction processing 42 for horizontal character strings, but the same applies to the confirmation and correction processing 43 for vertical character strings.

In the above embodiment, the function keys are mainly used to
I am trying to select the processing type.

For example, the processing type may be selected based on menu item hints from a pull-down menu or the like.

Furthermore, in the above embodiment, the processing device 10 performs confirmation and correction processing based on the input information of the second drawing input device 16, but by incorporating this function into the drawing input device 6, the host processing device Implementation that reduces the processing load is also possible.

〔Effect of the invention〕

As explained above, according to the present invention, a plurality of recognized characters and character cutout images from the original image are displayed on one screen at a time, and the result of single character recognition is checked and corrected. , operations for confirmation and correction are simplified, and the time required for it can be significantly shortened.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of the principle of the present invention. FIG. 2 is an example of recognized character data in one embodiment of the present invention. FIG. 3 is an example of a processing configuration according to an embodiment of the present invention. FIG. 4 is an example of a confirmation/correction screen in an embodiment of the present invention. FIG. 5 is an example of confirmation and correction according to an embodiment of the present invention. FIG. 6 is an example of horizontal character string confirmation and correction processing according to an embodiment of the present invention. FIG. 7 shows an example of the conventional method. In the figure, 10 is a processing device, 11 is a recognized character display processing section, 1
2 is a recognized character correction processing unit, 13 is a peripheral image display processing unit 514 which is a recognized character table, 15 is an image expansion memory, and 16 is a drawing input device. 17 is a scanner, 18 is a figure recognition device, 19 is a human power device, 20 is a keyboard, 21 is a tablet 22 is a primary recognition file, and 23 is a display. 24 is a display screen, 25 is a recognized character display area, 2
5 represents a character extraction image display area, 27 represents a peripheral image display area, and 28 represents a cursor.

Claims (1)

[Claims] In a drawing input system, characters extracted from an input drawing and recognized are displayed on a display, and the character recognition result confirmation/correction processing method is used to perform processing related to confirmation/correction of the character recognition results. Character display area (25) and character extraction image display area (2
6) A recognition method in which a plurality of sets are provided, and the recognized character result and the image to be recognized are displayed in correspondence with each other for each recognized character string in the recognized character display area and the character extraction image display area. The character display processing means (11) and the recognized character correction processing means (12) for updating the recognized character by screen correction in the recognized character display area, and the peripheral image containing the recognized character to be corrected are made to be the same. Peripheral image display processing means (
13) A character recognition result confirmation and correction processing method in drawing recognition processing.