JPH1011530A - Image input and output method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily obtain the correct result of recognition when a character part is extracted from an input document image and displayed after its recognition. SOLUTION: A document image is read by a scanner circuit 1 as the digital data and stored in a memory circuit 3. An area division circuit 4 decides a character part from an image part, and this decision result is stored in an area memory 5. At the same time, a recognition circuit 6 recognizes the characters of the character part and displays them on the screen of a display device 10. An information processing circuit 9 reconfigures the recognition result of the character part and displays both images and characters.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for inputting / outputting a document image, and more particularly to an image input / output device capable of recognizing a character portion in a document image and correcting the result.

[0002]

2. Description of the Related Art Conventionally, as this type of image input / output device, for example, a filing device capable of performing full-text search using character recognition is known. This is to file a pair of the character information and the document image that have been recognized as characters, or to file only the code information that has been recognized as characters. However, all of the above-mentioned code information is used as a keyword when searching. .

[0003] In the case of such an apparatus utilizing character recognition for full-text search, the recognition result does not necessarily have to be perfect. Even if there is a slight difference in the recognition result, there is no practical problem. In order to output to the application software and reuse it as code information, it is necessary to complete the document. Therefore, the difference is confirmed by comparing the result of character recognition with the document image and correcting it. .

[0004]

However, in the above-described conventional image input / output device, if an erroneous recognition is made, it can be found relatively easily. If the paragraph is complicated or the character part is arranged in a complicated manner, the order of performing the recognition process of the character part is often wrong, which breaks the context of the recognition result and restores it based on this It is a difficult task to do
In addition, when a large number of recognition results are used, there is a problem that this troublesome work must be repeated.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems. An image input / output method and a method for easily obtaining a correct recognition result when a character portion in an input document image is recognized. It is intended to provide a device.

[0006]

The image input / output device according to the present invention is configured as follows.

(1) Extraction means for extracting a character part from an input document image, display means for displaying the extracted character part, and re-recognition result of an arbitrary character part selected from the character parts And processing means to be configured.

(2) In the apparatus of the above (1), the processing means changes the order of recognition results of the selected plural character parts.

(3) In the apparatus according to the above (1) or (2), the processing means combines a plurality of selected character parts into one character part.

(4) In any one of the above-mentioned devices (1) to (3), the display means combines the area representing the character portion with the image and displays it.

(5) In any one of the above-mentioned devices (1) to (4), the display means displays the order of the recognition processing in the area representing the character portion.

(6) In the device of (4) or (5) above, the region representing the character portion is a rectangular region circumscribing the character portion.

An image input / output method according to the present invention is configured as follows.

(7) An extraction step of extracting a character part from the input document image, a display step of displaying the extracted character part, and a recognizing result of an arbitrary character part selected from the character part. And processing steps to be constituted.

(8) In the method of the above (7), in the processing step, the order of the recognition results of the selected plural character parts is changed.

(9) In the method of the above (7) or (8), in the processing step, a plurality of selected character portions are combined into one character portion.

(10) In any one of the above methods (7) to (9), the display step includes a step of combining an area representing a character portion with an image and displaying the image.

(11) In any one of the above methods (7) to (10), the displaying step includes a step of displaying the order of the recognition processing in the area representing the character portion.

(12) In the method of (10) or (11), a display step is provided in which the region representing the character portion is a rectangular region circumscribing the character portion.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to embodiments.

FIG. 1 is a block diagram showing the configuration of an image input / output device according to the present invention. In FIG. 1, 1 is a scanner circuit (input means) for taking in a document image, 2 is a memory control circuit for controlling a memory circuit (storage means) 3 for storing the document image, and 4 is a character portion according to the characteristics of the input document image. An area dividing circuit (extracting means) for dividing an area or the like, 5 an area memory for storing the area dividing result, 6
Extracts and compares the features of the image,
A recognition circuit (recognition means) for recognizing symbols, etc., 7 a recognition memory for storing the recognition result, 8 a dictionary memory, 9 an information processing circuit (processing means) for processing input information, and a plurality of arbitrary characters. A part is selected, and the recognition results of the selected character parts are reconstructed.

Reference numeral 10 denotes a display device (display means) such as a monitor for displaying a character portion or an image, 11 an external storage device such as an MO drive or a hard disk, and 12 a CPU or the like which controls the entire system of the image input / output device. And controls the entire system according to a program stored in the ROM 13. The ROM 13 stores a control program for processing according to the present invention, such as a flowchart described later.

Next, the operation will be described with reference to the flowchart of FIG. The operation of this flowchart is executed by the system control circuit 12.

First, when a document image reading instruction is given to the scanner circuit 1 under the control of the system control circuit 12, the scanner circuit 1 reads the document image as digital image data (hereinafter referred to as image data) (step S1). . The read image data is stored in the memory circuit 3 under the control of the memory control circuit 2 (step S2).

When the image data for one document image is stored in the memory circuit 3, the system control circuit 12
Instructs the memory control circuit 2 to output the read image data to the area dividing circuit 4 and gives an operation instruction to the area dividing circuit 4. As a result, the area dividing circuit 4 determines the type of the image in the input image data (step S3), and stores the result in the area memory 5 (step S4).

Here, as the area determination processing, for example, thinning processing is performed on input image data in order to increase the processing speed, or contour tracing is performed to detect continuity of black pixels. Then, a rectangle circumscribing the region where the black pixels are continuous is extracted, and the size of the rectangle and the proportion of the black pixels present in the rectangle are determined by the size of the region. Is determined.

In general, a character image may be written vertically or horizontally. The distance between the character and the rectangular area determined to be the character is checked. If the distance between the character image and the horizontal direction is small, the character image is written horizontally and vertically. If the distance is short, it can be determined that the document is written vertically.

In the case of a photograph or the like, the size of the rectangle is a certain size or more and the number of black pixels is large.
In the case of the table, the rectangles are arranged neatly, and the ratio of the number of black pixels in the rectangle is low, so that it can be easily determined.

In the present embodiment, the region is divided as described above, but it goes without saying that other algorithms may be used. In short, in the present embodiment, it is sufficient that at least the character image portion and the non-character portion can be identified.

In some cases, the user may view the display screen of each area, set the area using a pointing device, and specify or correct the type.

As a result, the area dividing circuit 4 constructs a table as shown in FIG. 3 in the area memory 5, for example.

In the table shown in FIG. 3, the first field is an area number assigned to each area, and the second field is information (coordinates of the upper left corner and the lower right corner) for specifying the area (rectangle). . In the third field, the type of each area is stored. In the present embodiment, “text” indicating a character image and “non-text” indicating a non-character such as a photographic image are used. I have.

Next, when the area dividing process by the above-described area dividing circuit 4 is completed, the system control circuit 12 refers to the area memory 5 and reads out the image data of the area determined as the text area from the memory circuit 3 ( Step S5), the data is output to the recognition circuit 6. At this time, the area dividing circuit 4 simply bypasses.

The recognition circuit 6 cuts out a character image from the given image data, extracts feature information from the cut-out character image, and compares it with the standard feature information in the dictionary memory 8 (step S6). Of character recognition processing.

In general, in such a character recognition process, the distance between feature information (feature vector) obtained from a character image to be recognized and standard feature information stored in a recognition dictionary (dictionary memory 8) is determined. The character (character code) having the standard feature information having the shortest distance is determined as the first candidate, and is output as the second candidate, the third candidate, and so on.
At this time, the maximum number of candidates is set to N, and no more candidates are output.

Next, the recognition circuit 6 outputs the first candidate and other candidates obtained by the above-described recognition processing to the recognition memory 7. At this time, the system control circuit 12 transfers the position information of the recognized character image (information indicating the position of the character in the document image) to the recognition circuit 6 and stores the information in the recognition memory 7 ( Step S7). As a result, a table as shown in FIG.

In the table shown in FIG. 4, information indicating which area number (see FIG. 3) the character is is in the first field, the position information of the character image in the second field, and the character code in the third field. (A first candidate, a second candidate,... With candidate characters obtained as a result of recognition) are respectively stored.

Although the size of each character is assumed to be substantially constant here, when recognizing a document image having characters of different sizes, two pieces of coordinate position information in the table of FIG. Become. That is, for example, the coordinate data of the upper left corner and the lower right corner.

Next, as described above, when the character recognition for the character area in the document image is completed, the bitmap data for the area determined to be other than the character in the input document image is replaced by the bit map data for the character recognized part. Is the recognition memory 7
The contents of the table stored in the external storage device 1
File 1 In addition, the contents of the table stored in the area memory 5 indicating the type and position of the document image are also simultaneously filed (step S8).

Here, when there are a plurality of document images,
For example, it is determined whether or not there is a document image by a document placement sensor of the automatic document feeder provided in the scanner circuit 1, and if so, the operation described above is repeated (step S9).

When an instruction from a keyboard or the like (not shown) is received and character code information recognized by another application or the like is used as plain text, an image arbitrarily selected from the external storage device 11 is read, and the read image is stored in the area. The image is displayed on the display device 10 based on the division result (step S10). The filing process and the display process are performed by the information processing circuit 9 under the control of the system control circuit 12.

Next, the display processing will be described. Here, for simplicity of explanation, it is assumed that the display device 10 in the present embodiment is a general-purpose monitor, and the information processing circuit 9 includes a buffer memory for storing a display image.

A non-character image (such as a photograph) is directly developed in the above buffer memory, and font data is extracted from a font memory (not shown) based on the first candidate character code as a result of recognition for the character image. A character pattern is generated and developed at a corresponding position according to the position information of the table shown in FIG.

At this time, since the type of each area is stored in the third field of the table of FIG. 3, only the text area has the area number stored in the first field and the area stored in the second field. According to the information specifying the (rectangle), a number indicating the order in which the recognition processing was performed is added to the image developed in the buffer memory,
After adding a frame representing the character image area, the display device 10
To be displayed.

FIG. 5 shows a specific example of the relationship between a document image and a display image. In the figure, reference numeral 50 denotes a document image (or image data stored in the memory circuit 3). As shown in the figure, the document image is a mixture of a photographic image and characters.

Reference numeral 51 denotes the display content of the display device 10, and reference numeral 52 denotes a plain text that can be used at this time. The context is shifted because the order of the recognition processing is incorrect.

Therefore, a keyboard (not shown) or the like is used to
By instructing to change the order of the displayed recognition processing as in the display contents of 3, the information processing circuit 9 changes the order of the character codes stored in the external storage device 11 under the control of the system control circuit 12. Then, a plain text as shown in 54 is formed and stored in the external storage device 11 as a separate file.

Here, when changing the order of the character codes, there are cases where it is desired to change the insertion position of the entire table, for example, when a table (not shown) includes a plurality of character portions. , A plurality of character parts in the table can be combined and treated as one character part, and the number of character parts to be handled can be reduced.

In this way, by reading another file stored in the external storage device 11 from another application, the character code information whose context has been correctly corrected is used as plain text.

A more detailed description will be given with reference to FIG.

First, regarding the numbers added to the display contents of the display device 10 and indicating the order in which the recognition processing has been performed, the numbers are created as follows.

FIG. 6A) shows 5 corresponding to the display content of 51.
The table contents constructed in the area memory are shown, but the area number of the text area only is selected by excluding the non-text area from the area number of the first field,
A new area number is set such that there is no empty number while preserving the respective order relations, and a correspondence table as shown in FIG. 6B) is created.

In this correspondence table, the first field is the area number of only the text area in the first field of FIG. 6A, and the second field is the above-mentioned new area number. This corresponds to a number indicating the order in which the processing was performed.

Therefore, when the order of the recognition processing is changed and the context is corrected as described in the display contents of 53,
For example, a new area number indicating the order in which the recognition processing was performed by giving an instruction using a keyboard (not shown) or the like is shown in FIG.
This can be done by modifying as shown in c). Here, the order of the second and third recognition processes is reversed, but the area number of the first field corresponding to the second field of the correspondence table is referred to by referring to the table of FIG. The character code corresponding to the number can be extracted. In this way, the character codes are read from the external storage device 11 in the order of the corrected new area numbers,
A plain text as shown at 54 will be formed. A case where a plurality of character parts in the above-described table are combined and treated as one character part will be described with reference to FIG.

FIG. 7A shows the contents of the table constructed in the area memory of area 5, and the character parts corresponding to the area numbers 4 to 7 are the character parts in the table.

FIG. 7B) shows the correspondence table before correction, and FIG. 7C) shows the correspondence table after correction.
The value of the second field in the character portion corresponding to .about.7 is changed equally, and the order of the second field in the character portion corresponding to 2 of the first field is reversed. This indicates that the character part in the table is inserted between the character parts 1 and 2 of the first field all together.

In this way, the value of the second field can be made equal for a plurality of character parts in the table, and the handling can be linked.

However, it is conceivable that the order of the character parts to be linked is changed, and the third field is extended to the correspondence table shown in FIG. 7C) to describe the order of the recognition processing of the character parts in the table. The order is changed by changing the order of the extensions.

As described above, for example, when it is desired to use the filed character recognition result, it is possible to easily recognize whether or not the recognition result of the character part is in context before moving to another application. If it is not satisfactory, the recognition result can be corrected appropriately using a keyboard (not shown) or the like.

In the above example, a character pattern is generated and displayed based on the character code of the recognition result. However, it may be displayed as it is by filing the bit map of all images. However, also in this case, a number indicating the order in which the recognition processing has been performed is added, and a frame indicating a character image area is added for each character portion.

In the above example, the order in which the recognition processing was performed is represented by a numeral, and the character image area is surrounded by a frame. However, the present invention is not limited to this, and alphabets and the like are used to represent the order. Alternatively, shading or display color may be changed instead of the frame. That is, any display form may be used as long as it can be distinguished.

In the above example, the plain text is stored as a separate file after the context is corrected by displaying and confirming the image based on the area division result. The contents of each table may be modified by creating and storing the correspondence table.

Further, in the above example, an example in which an image based on the region division result is displayed has been described, but the image may be printed by a print output device. In this case, by outputting a plurality of sheets, it is possible to perform correction processing collectively later.

Further, the apparatus shown in this embodiment may be applied to a general-purpose information processing apparatus. The device configuration in this case is, for example, as shown in FIG.

Referring to FIG. 8, reference numeral 100 denotes a CPU for controlling the apparatus or the entire system; 101, a ROM for storing a boot program and a BIOS;
The image storage area 103 corresponding to the memory circuit 3 of FIG. 1, the area type area 104 corresponding to the area memory 5, the recognition memory 7
The following data or programs are developed or secured, including the recognition result area 105 of 105 corresponding to.

That is, 106 is a program corresponding to a series of processing procedures, 107 is an output buffer area used for printing, and 108 is an operating system (OS) for controlling the entire apparatus including the control program 106.

Reference numeral 109 denotes an image scanner for reading a document image, and 110 denotes a large-capacity storage device such as a hard disk device or a magneto-optical disk device, in which an operating system 108 and the like are also stored. An apparatus 111 for reading data from a portable storage medium is a floppy disk apparatus having a disk drive in the figure. A floppy disk set in the floppy disk device 111;
Alternatively, the control program 106 is stored in one of the storage devices 110 such as a hard disk device.

Reference numeral 112 denotes a VRAM for expanding a bitmap image to be displayed on a display screen, and 113 denotes a VR.
A display for displaying the bitmap image developed on the AM 112, a keyboard 114 for inputting various information for correcting the recognition result, and a 115 for designating a desired position on the display screen of the display 113 And a pointing device 116 for selecting one of various menus.

The operation of the device having the above configuration is almost the same as that of the embodiment described above, but the processing corresponding to the memory control circuit 2 and the area dividing circuit 4 is the same as that of the CP.
U100 is performed according to the control program 106.

As described above, the present invention can also be achieved by externally supplying a program for implementing the above-described processing to a general-purpose information processing apparatus or system.

Therefore, the present invention may be applied to a system including a plurality of devices such as a host computer, an interface, and a printer, or to an apparatus including a single device such as a copying machine.

It is needless to say that the present invention can be attained by supplying a program stored in a storage medium to a system or an apparatus.

[0073]

As described above, according to the present invention,
When a character portion in an input document image is recognized, a more correct recognition result can be easily obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image input / output device according to the present invention.

FIG. 2 is a flowchart showing the operation of the apparatus of FIG. 1;

FIG. 3 is a view showing contents of a table constructed in an area memory of FIG. 1;

FIG. 4 is a diagram showing contents of a table constructed in the recognition memory of FIG. 1;

FIG. 5 is an explanatory diagram illustrating a relationship between a document image and a display image according to the embodiment;

FIG. 6 is a diagram illustrating a relationship between table contents and a correspondence table in the embodiment.

FIG. 7 is a diagram illustrating a relationship between table contents and a correspondence table according to the embodiment;

FIG. 8 is a block diagram showing another configuration of the device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 scanner circuit (input means) 2 memory control circuit 3 memory circuit (storage means) 4 area dividing circuit (extraction means) 5 area memory 6 recognition circuit (recognition means) 7 recognition memory 8 dictionary memory 9 information processing circuit (processing means) Reference Signs List 10 display device (display means) 11 external storage device 12 system control circuit 13 ROM

Claims (12)

[Claims] 1. An extracting means for extracting a character part from an input document image, a display means for displaying the extracted character part,
Processing means for reconstructing a recognition result of an arbitrary character portion selected from the character portions. 2. The method according to claim 1, wherein the processing unit changes the order of the recognition results of the plurality of selected character parts.
The image input / output device according to the above. 3. The method according to claim 1, wherein the processing unit combines the plurality of selected character parts into one character part.
Or the image input / output device according to 2. 4. The image input / output apparatus according to claim 1, wherein the display unit combines and displays an area representing a character portion with the image. 5. The image input / output apparatus according to claim 1, wherein the display means displays the order of the recognition processing in an area representing a character portion. 6. The image input / output device according to claim 4, wherein the region representing the character portion is a rectangular region circumscribing the character portion. 7. An extracting step of extracting a character part from an input document image, a displaying step of displaying the extracted character part,
Reconstructing a recognition result of an arbitrary character portion selected from the character portions. 8. The processing step according to claim 7, wherein the order of the recognition results of the selected plurality of character parts is changed.
Image input / output method described. 9. The processing step of combining a plurality of selected character parts into one character part.
Or the image input / output method according to 8. 10. The image input / output method according to claim 7, wherein the displaying step includes a step of combining and displaying an area representing a character portion with the image. 11. The image input / output method according to claim 7, wherein the displaying step includes a step of displaying an order of the recognition processing in an area representing a character portion. 12. The image input / output method according to claim 10, further comprising a display step of indicating that the region representing the character portion is a rectangular region circumscribing the character portion.