JPH096974A - Image processing method and apparatus - Google Patents

Abstract

(57) [Summary] [Purpose] In image information in which image information areas of various attributes such as a text area and a table area coexist, whether or not to output each attribute can be controlled by a simple operation. [Structure] Image information is input (step S201), an area is divided from the input image (step S202), and an identification number for identifying an attribute is added to each area and stored as rectangular data (step S203, FIG. 5) change the output flag in the stored rectangular data according to the output target selection process (step S301) (step S302),
The output is controlled according to this rectangular data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic device such as an OCR (optical character recognition) device, a copying machine, a facsimile, etc., particularly a text area, a table area, a graphic area, a text area in a chart, etc. for an input image. The present invention relates to an image processing method and apparatus that can be divided into areas.

[0002]

2. Description of the Related Art In a conventional image processing apparatus, as an automatic area dividing means, there is one that divides into a text area and an area other than text and performs output processing. Further, there is a method in which a ruled line is extracted for the table area and the text in the area surrounded by the ruled line is extracted.

[0003]

However, in the above-mentioned prior art, when a table is included in the text, the text in the text and the text in the table are not removed even if the text in the table is extracted as the text in the table. Since they are always output together, the text and table are displayed at the same time even when recognition is performed, and the recognition results cannot be separated, and the text in the table is mixed in the text.

Therefore, even if only the in-table text area is desired, in order to obtain it, it is necessary to delete other text areas after output or to specify only the table area in advance. There was a drawback.

[0005]

Means for Solving the Invention

[0006]

In order to solve the above-mentioned conventional problems, the present invention extracts a region in which an image exists based on input image information and corresponds to a non-text discriminated from the extracted region. The text area included in the area to be detected is detected, and when the selective output of the text included in the non-text area is instructed, the image of the text area included in the detected non-text area is output. There is provided an image processing method and apparatus for controlling so.

Further, in order to solve the above-mentioned conventional problems, the present invention preferably, when the selective output of the text included in the non-text area is not instructed, the text included in the non-text area is preferably instructed. Control not to output the image of the area.

In order to solve the above conventional problems, the present invention preferably sets the non-text area to an image area in which a table is drawn.

In order to solve the above-mentioned conventional problems, the present invention preferably sets the non-text area to an image area in which a graphic is drawn.

In order to solve the above-mentioned conventional problems, the present invention preferably provides the non-text selective output instruction,
Done on screen.

In order to solve the above-mentioned conventional problems, the present invention preferably gives an instruction for the selective output of the text for each same attribute.

In order to solve the above-mentioned conventional problems, the present invention is preferably such that the input image information is a mixture of images having a plurality of types of attributes.

In order to solve the above-mentioned conventional problems, the present invention stores, for each area, an area in which an image extracted from input image information exists and an image attribute of each area, and indicates a desired attribute, Provided is an image processing method and apparatus for controlling an area of the designated attribute according to the stored area information and outputting image information of the area determined to be the area of the designated attribute. .

In order to solve the above-mentioned conventional problems, the present invention preferably gives an instruction as to whether or not to output the area having the instructed attribute, and controls the output of the image information according to the instruction. To do.

In order to solve the above conventional problems, the present invention preferably sets the attribute of the image to be stored as information indicating that the image in the area is text.

In order to solve the above conventional problems, the present invention preferably sets the attribute of the image to be stored as information indicating that the image of the area is a figure.

In order to solve the above conventional problems, the present invention preferably sets the attribute of the image to be stored as information indicating that the image of the area is a table.

Further, in order to solve the above-mentioned conventional problems, in the present invention, the attribute of the image to be stored is preferably information indicating that the image in the area is text in the figure.

In order to solve the above-mentioned conventional problems, the present invention preferably sets the attribute of the image to be stored as information indicating that the image of the area is the in-table text.

In order to solve the above-mentioned conventional problems, the present invention preferably gives the instruction on the screen.

In order to solve the above conventional problems, the present invention preferably gives the output control instruction on a screen.

In order to solve the above conventional problems, the present invention presents the attributes of images existing in the image information to be processed, and outputs the image information of each attribute for the presented attributes. There is provided an image processing method and apparatus for instructing whether or not to output, and storing output control information of an image based on the instruction.

Further, in order to solve the above-mentioned conventional problems, the present invention preferably relates to an area where an image included in the image information exists, and area information including position information of each area and attribute information of the image of the area. And the attributes presented are:
The attribute is stored as the area information.

In order to solve the above conventional problems, the present invention preferably includes the output control information of the image in the area information.

Further, in order to solve the above-mentioned conventional problems, in the present invention, preferably, the instruction is on a display screen.

[0026]

【Example】

(First Embodiment) FIG. 1 is a block diagram showing the arrangement of an image processing apparatus according to the first embodiment. In FIG. 1, 1
Reference numeral 01 denotes a scanner that irradiates an image original with light and reads the reflected light to convert it into an electric signal. Reference numeral 102 converts the electric signal obtained by the scanner 101 into a binary digital electric signal and transmits it to other device components. A scanner interface circuit for inputting, 103 is a pointing device (mouse or the like) for inputting desired coordinates on the window of the display 108, a keyboard 113 is a keyboard for inputting various control commands and character codes, and 104 is a pointing device. An interface circuit for receiving signals from the 103 and the keyboard 113 and transmitting them to other device constituent elements, 105 is for executing control of the entire device and character cutting processing and recognition processing according to a control program in the ROM 106. CPU, 106 is shown in the flow chart described later. Processing control program of CPU105 executes, ROM that stores various processing programs and font data, 107 is used as such a work area for the character image expansion and character recognition processing, also,
This is a RAM for storing reference parameters and the like used when determining the attributes of the image area. Reference numeral 108 is a display for displaying an input image and a recognition result, which is a CRT or a liquid crystal display. 109 is a display interface circuit. On the display 108,
V stored in a predetermined address area of the RAM 107
Display an image of the RAM area. Reference numeral 110 is an external storage device such as a hard disk in which registered data is stored, and 111 is its interface. And 112
Is a bus for connecting each device component. Further, the printer 115 outputs information or a recognition result that is to be output based on the output control of the present invention, and is composed of an LBP or an inkjet printer. 114
Is an interface with the printer 115. 11
Reference numeral 6 denotes a communication interface, which controls data communication through a public line or the like, and also inputs an image and outputs a result.

Next, the process flow will be described with reference to the flow chart of FIG. 2 and FIGS. 3, 4, 5, 6, 7, 8 and 9.

First, an embodiment of extracting various areas will be described.

In step S201, the image of the document to be recognized is read by the scanner 101, converted into binary image data, and stored in the RAM 107.

The input image obtained by the scanner 101 is analyzed and area division processing is performed in S202, and various areas having different attributes such as a text area, a figure area, and a table area are defined in advance for each attribute. Extract according to the conditions. Various attributes (text, figures, tables, etc.) in various areas
The identification code corresponding to is added to the rectangular data representing the area extracted in S202 as a data attribute.

Further, in consideration of the case where there is a character included in the non-text area estimated to be the table area or the graphic area in S203, the extraction processing of the text areas in those areas is performed. The process of extracting the text in the non-text area is repeated for all the non-text areas extracted in S202. The text in the non-text area extracted here indicates that it is the text area extracted from the non-text (table or figure) as "text in table" or "text in figure" as the attribute of each area. An identification code different from that of the text area of the body is added to the rectangular data. In FIG. 5, the RAM 10 having the rectangular shape extracted in the processing up to S203 from the image information of the document input in S201.
7 shows the format of data when stored in 7.
(A) is a format of data in units of rectangles (text area, table area, figure area, in-table text area, in-figure text area, etc.), and a start point and an end point coordinate for specifying each rectangle and the rectangle Attribute, the recognition order of the rectangles in the plurality of rectangles extracted from the image input in S201, and the output flag described in detail later are stored. As for the identification number of the attribute stored here, as shown in (b), when the attribute is "body text", "1" is stored in the attribute storage area in the rectangular data, and the attribute is "in the table". text"
If it is, the attribute storage area in the rectangular data is “2”.
Is stored, and the attribute of each rectangle can be identified by the identification number as shown in (b).

In step S204, the output flag is set to 0 for all the areas and initialized, and all the rectangular data is temporarily held in the memory (S205). If the output flag is "0", it is "valid", that is, in the selected state, and is displayed when the display of the document is instructed. On the contrary, if the output flag is "1", it is "invalid" and this rectangle is not displayed.

Next, an example of selecting a text area will be described by taking the document of FIG. 6 as an example.

In the first processing after the first area division,
Since the output flag has been initialized to 0 in S204,
Both the body text area and the in-table text area are valid, and both areas are displayed on the display 108 (FIG. 6).

Next, in order to select a text having a desired attribute from a plurality of rectangular areas displayed in FIG. 6 with a pointing device or the like, a text having a desired attribute such as "body text" or "text in table" is selected. A soft switch is selected (S301). These soft switches
When the soft switch 61 called "character recognition target" as shown in FIG. 6 is clicked with the mouse 103, the soft switch showing the attribute included in the displayed image is shown as 81 in FIG. Is displayed.

In this example, "text" and "text in table"
Every time the switch is clicked with the mouse, it becomes "valid", so that it becomes "valid" by clicking the switch 82 "invalid" in the column I try to reverse "invalid". Details of the output flag changing process (S302) of the output target data are shown in the flowchart of FIG. 4 and will be described below. This process is for mouse 1
It starts from the judgment that the area clicked by 03 is the area of the switch indicating "invalid" or "valid".

Here, the valid / invalid switch 83 of the text is used.
A description will be given of the processing when both the valid / invalid switch 82 for the in-table text are clicked by the mouse 103.

First, the rectangular data in the RAM 107 is searched for the attribute "1", that is, the rectangular data which is the body text, and when the body valid / invalid switch 83 is instructed to "valid" (S401). The output flag of the searched rectangular data is set to "0" (S402), and when the valid / invalid switch of the text is instructed to "invalid" (S401), the output flag of the searched rectangular data is set. It is set to "1" (S403). Similarly to the body text, the in-table text searches the rectangular data in the RAM 107 for rectangular data whose attribute is “2”, that is, in-table text, and how the in-table text valid / invalid switch 82 is set. Is instructed (S
404), and the output flag of the rectangular data searched accordingly is set to "0" (S405) or "1" (S406).
Set to.

The setting process of these output flags is valid.
This is repeated for the rectangular data having the attribute whose change is instructed by the invalid switch.

Then, in S303, the text area whose output flag is 0 is output.

The examples of FIGS. 7, 8 and 9 are examples in which various situations are selected as the character recognition target. In FIG. 7, the body recognition text and the in-table text are initially set without changing the character recognition target. Both are valid examples, and all text rectangles are displayed with a frame to indicate that they are valid (however, the dotted line indicates a non-text area), that is, "text" and " It is shown that both "text in table" are output, and when the "recognition" button 71 is clicked, all the text surrounded by the solid line frame is recognized.

FIG. 8 shows an example in which the "body" is set to "valid" and the "text in table" is set to "invalid". A frame is displayed in the "body" area to show that it is valid. That is, only the "text" indicates that output is instructed,
When the "recognition" button 84 is clicked, the "body" surrounded by the solid line frame is recognized.

FIG. 9 shows an example in which the "text" is "invalid" and the "text in table" is "valid", and a frame is displayed in the area of "text in the table" and it is valid. Indicates
That is, only the "text in the table" indicates that the output is instructed, and when the "recognition" button 91 is clicked, the "text in the table" is recognized.

(Second Embodiment) In this embodiment, an example in which area division and character recognition are executed as a series of operations will be described. Then, at that time, in the output target selection means, the contents of the output target which have been changed once can be stored in the holding file in the output target in the RAM 107, and the stored contents can be executed as an initial value at the next operation. Is. The output target holding file stores attributes and information indicating whether the output flag of each attribute is “0” or “1”.

An example of this processing is shown in the flowchart of FIG. 10. First, the output target is read from the output target holding file in the RAM 107 (S1001), and the output target is selected and initialized according to the stored contents (S1002). If both "body" and "text in table" are enabled in the output target holding file by initialization, the output flags of both text data are initialized to 0.

Next, when the output target is selected by the pointing device or the like, the output flag of the output target data is changed (S1004). This process is similar to the process shown in the flowchart of FIG.

The change at this time is stored in the output target holding file (S1005), and after these processes, the text area in which the output flag is 0 is output as the character recognition target (S1006).

When the output target is not selected by the pointing device or the like, the text area in which the output flag is 0 is output as it is as the character recognition target in the state set in S1002 (S1006).

(Third Embodiment) In this embodiment, an example in which it is possible to select whether or not to output not only in a table but also in a figure or text in a frame will be described. FIG. 11 shows the text 11 in the figure extracted from the figure area 1100.
This is an example in which whether or not 01 is output is added to the output target selection means, and a switch 1102 for text in the figure is also displayed as a valid / invalid switch. Since "text in drawing" is set to "valid" by this switch, a frame is displayed in the text area 1001 in the figure to indicate to the user that the output is instructed to be valid, and "recognition" is performed. When the button 1103 “” is clicked, the “text in the figure” is also recognized.

In the first to third embodiments, an example in which the area is automatically divided has been described. However, a user manually specifies a rectangle and an attribute in the input image, or the area is automatically specified. The extraction is performed automatically, but the attribute is specified manually by the user, or the area is automatically divided as in the present embodiment, but the area and attribute are corrected manually by the user. As for the example, as in the example shown in FIG. 5, the present application is of course applicable to data having area data, attribute data, and an output flag for each rectangle or attribute.

In the first to third embodiments, the process of selecting the output target as the character recognition target has been described as an example. Therefore, the attribute of the selection target region is only the text type. Obviously, if the rectangular data having the attribute information and the output flag and the output selection technique are used, the target to be selected as the output target can be metallic.

By doing so, an image obtained by removing only the image from the image in which the text, the image and the table are mixed,
It is possible to output images in various combinations, such as an image in which only text is removed, or an image in which only a table is included, and it can be used in the field of image editing.

[0053]

As described above, according to the present invention,
From the input image information, extract the area where the image exists,
The text area included in the area corresponding to the non-text identified from the extracted area is detected, and when the selective output of the text included in the non-text area is instructed, the detected non-text is detected. By controlling the output of the image of the text area included in the text area, it is possible to easily instruct to output the text in the non-text extracted from the input image.

As described above, according to the present invention, when the selective output of the text included in the non-text area is not instructed, the image of the text area included in the non-text area is output. By controlling not to do so, it is possible to easily prevent the text in the non-text from being output.

As described above, according to the present invention, the non-text area is an image area in which a table is drawn, so that the output of the text in the table can be easily controlled.

As described above, according to the present invention, the non-text area is an image area in which a graphic is drawn, so that the output of the text in the graphic can be easily controlled.

As described above, according to the present invention, the instruction for the selective output of the text is given on the screen, so that the operability can be further improved.

As described above, according to the present invention, the instruction of the selective output of the text can be performed for each attribute of the selective output by instructing the same attribute, so that the operability can be further improved.

As described above, according to the present invention, it is possible to process a mixed image of a plurality of types of attributes, so that it is possible to easily process more complicated image information.

As described above, according to the present invention, the area where the image extracted from the input image information and the attribute of each area image are stored for each area, and the desired attribute is designated, Since the area of the designated attribute is determined according to the stored area information and the image information of the area determined to be the area of the designated attribute is controlled to be output, each area extracted from the image is controlled. Can be easily instructed.

As described above, according to the present invention, by instructing whether or not to output the area having the instructed attribute, and outputting the image information in accordance with the instruction, the image information is output for each attribute. The output can be easily controlled.

As described above, according to the present invention, the attribute of the image to be stored is information indicating that the image in the area is text, so that the output control for the text becomes easy.

As described above, according to the present invention, the attribute of the image to be stored is information indicating that the image in the area is a figure, so that the output control for the figure is facilitated.

As described above, according to the present invention, the attribute of the image to be stored is information indicating that the image in the area is a table, so that control of output to the table becomes easy.

As described above, according to the present invention, the attribute of the image to be stored is the information indicating that the image of the area is the text in the figure, so that the output control for the text in the figure can be controlled. It will be easier.

As described above, according to the present invention, the attribute of the image to be stored is information indicating that the image of the area is the in-table text, so that the output control for the in-table text can be controlled. It will be easier.

As described above, according to the present invention, the instruction is given on the screen, so that the operability is further improved.

As described above, according to the present invention, the operability is further improved by instructing the output control instruction on the screen.

As described above, according to the present invention, the attribute of the image existing in the image information to be processed is presented, and the image information of each attribute is output for the presented attribute. By instructing whether to output or not and storing the image output control information based on the instruction, the input and instruction of the image output control information can be easily performed.

As described above, according to the present invention, regarding the area where the image included in the image information exists, the area information including the position information of each area and the attribute information of the image of the area is stored, and By setting the attribute to be presented as the attribute stored as the area information, it becomes easy to manage the output for each image area.

As described above, according to the present invention, the output control information of the image is included in the area information, whereby the output management becomes easier.

As described above, according to the present invention, by providing the instruction on the display screen, the operability is further improved.

[Brief description of drawings]

FIG. 1 is a block diagram of a first embodiment.

FIG. 2 is a flowchart of detecting various areas according to the first embodiment.

FIG. 3 is a flowchart of selecting an output target according to the first embodiment.

FIG. 4 is a flowchart for changing an output flag of output target data according to the first embodiment.

FIG. 5 is an example of a structure of rectangular data according to the first embodiment.

FIG. 6 is an example of a document image and a GUI (graphical user interface) of the first embodiment.

FIG. 7 is an example of output of all areas in the first embodiment.

FIG. 8 is an example of output of a text-only area according to the first embodiment.

FIG. 9 is an example of output of a text-only area in a table according to the first embodiment.

FIG. 10 is a flowchart of the second embodiment.

FIG. 11 is an example of output of the third embodiment.

Claims (50)

[Claims] 1. A region in which an image exists is extracted from input image information, a text region included in a region corresponding to a non-text identified from the extracted region is detected, and a non-text region is detected. An image processing method, comprising: controlling to output an image of a text area included in the detected non-text area when the selective output of the included text is instructed. 2. When the selective output of the text included in the non-text area is not instructed, the image of the text area included in the non-text area is controlled not to be output. The image processing method according to item 1. 3. The image processing method according to claim 1, wherein the non-text area is an image area in which a table is drawn. 4. The image processing method according to claim 1, wherein the non-text area is an image area in which a graphic is drawn. 5. The image processing method according to claim 1, wherein the instruction to select and output the text is given on a screen. 6. The image processing method according to claim 1, wherein the instruction to select and output the text is issued for each same attribute. 7. The image processing method according to claim 1, wherein the input image information is a mixture of images of a plurality of types of attributes. 8. An area in which an image extracted from input image information and an attribute of an image of each area are stored for each area, a desired attribute is designated, and the area of the designated attribute is stored in the stored area. The image processing method is characterized in that the image information of the area determined to be the area having the designated attribute is output. 9. The image processing method according to claim 8, further comprising instructing whether or not to output the area having the instructed attribute, and controlling the output of the image information in accordance with the instruction. . 10. The image processing method according to claim 8, wherein the attribute of the image to be stored is information indicating that the image in the area is text. 11. The image processing method according to claim 8, wherein the attribute of the image to be stored is information indicating that the image in the area is a figure. 12. The image processing method according to claim 8, wherein the attribute of the image to be stored is information indicating that the image of the area is a table. 13. The image processing method according to claim 8, wherein the attribute of the image to be stored is information indicating that the image in the area is text in the figure. 14. The image processing method according to claim 8, wherein the attribute of the image to be stored is information indicating that the image in the area is in-table text. 15. The image processing method according to claim 8, wherein the instruction is given on a screen. 16. The image processing method according to claim 8, wherein the output control instruction is given on a screen. 17. Instructing an attribute of an image existing in image information to be processed, instructing whether or not to output image information of each attribute for the instructed attribute, an image based on the instruction An image processing method characterized by storing the output control information of the above. 18. Regarding the area where the image included in the image information exists, area information including position information of each area and attribute information of the image of the area is stored, and the presented attribute is stored as the area information. 18. The image processing method according to claim 17, wherein the attribute is a specified attribute. 19. The image processing method according to claim 17, wherein the output control information of the image is included in the area information. 20. The image processing method according to claim 17, wherein the instruction is on a display screen. 21. The image processing method according to claim 17, wherein the attribute of the image to be presented is text. 22. The image processing method according to claim 17, wherein the attribute of the image to be presented is a table. 23. The image processing method according to claim 17, wherein the attribute of the image to be presented is a graphic. 24. The image processing method according to claim 17, wherein the attribute of the image to be presented is in-table text. 25. The image processing method according to claim 17, wherein the attribute of the image to be presented is in-text. 26. An image inputting means for inputting an image, an area extracting means for extracting an area where an image exists based on the image information input by the image inputting means, and an area extracted by the area extracting means are identified. In the non-text area, the text detection means for detecting the text area included in the area corresponding to the non-text area, and the selective output of the text included in the non-text area An image processing apparatus, comprising: output control means for controlling to output an image of a text area included in the non-text area detected by the text detection means. 27. When the selective output of the text included in the non-text area is not instructed, the output control means controls not to output the image of the text area included in the non-text area. The image processing apparatus according to claim 26, wherein: 28. The image processing apparatus according to claim 26, wherein the non-text area is an image area in which a table is drawn. 29. The image processing apparatus according to claim 26, wherein the non-text area is an image area in which a graphic is drawn. 30. The image processing apparatus according to claim 26, wherein the instruction to select and output the text is given on a screen. 31. The image processing apparatus according to claim 26, wherein the instruction to select and output the text is issued for each same attribute. 32. The image processing apparatus according to claim 26, wherein the image information input by the image input means is a mixture of images of a plurality of types of attributes. 33. Image input means for inputting an image, and area information storage means for storing, for each area, an area in which an image extracted from the image information input by the image input means exists and an attribute of the image of each area. An instruction means for instructing a desired attribute, and an area of the attribute instructed by the instructing means is determined according to area information stored in the area information storage means,
An image processing apparatus, comprising: output control means for controlling to output image information of an area determined to be the area of the designated attribute. 34. An output control instructing means for instructing whether or not to output the area having the attribute instructed by the instructing means, and controlling the output by the output control means in accordance with the instruction of the output control instructing means. 34. The image processing device according to claim 33, wherein 35. The image processing apparatus according to claim 33, wherein the attribute of the image stored in the area information storage means is information indicating that the image in the area is text. 36. The image processing apparatus according to claim 33, wherein the attribute of the image stored in the area information storage means is information indicating that the image of the area is a figure. 37. The image processing apparatus according to claim 33, wherein the attribute of the image stored in the area information storage means is information indicating that the image of the area is a table. 38. The image processing apparatus according to claim 33, wherein the attribute of the image stored in the area information storage means is information indicating that the image of the area is text in the figure. 39. The image processing apparatus according to claim 33, wherein the attribute of the image stored in the area information storage means is information indicating that the image of the area is the in-table text. 40. The image processing apparatus according to claim 33, wherein the instructing means is means for instructing on a screen. 41. The image processing apparatus according to claim 33, wherein the output control instructing means is means for instructing on the screen. 42. Attribute presenting means for presenting the attribute of the image existing in the image information to be processed, and whether or not to output the image information of each attribute for the attribute presented by the attribute presenting means. An image processing apparatus comprising: an instruction unit for instructing, and a storage control unit for controlling to store output control information of an image based on the instruction of the instruction unit. 43. Regarding an area in which an image included in the image information exists, area information storage means for storing area information of each area and attribute information of the image of the area is provided, and the attribute presentation means is the area information. The image processing apparatus according to claim 42, characterized in that the attributes stored in the storage means are presented. 44. The image processing apparatus according to claim 42, wherein the storage control unit controls to store the output control information in the area information storage unit. 45. The image processing apparatus according to claim 42, wherein the instruction unit gives an instruction on a display screen. 46. The attribute of the image presented by the attribute presenting means is text.
An image processing apparatus according to claim 1. 47. The image processing apparatus according to claim 42, wherein the attribute of the image presented by the attribute presenting means is a table. 48. The image processing apparatus according to claim 42, wherein the attribute of the image presented by the attribute presenting means is a graphic. 49. The image processing apparatus according to claim 42, wherein the attribute of the image presented by the attribute presenting means is in-table text. 50. The image processing apparatus according to claim 42, wherein the attribute of the image presented by the attribute presenting means is in-text.