JP6707903B2 - Output device and computer program - Google Patents

Description

The present specification relates to a technique of outputting image data.

Conventionally, a technique of outputting new image data by using the image data generated by processing the image data has been used. For example, a technique has been proposed in which a character size is estimated and new image data is output using image data in which the image resolution is adjusted according to the estimated character size.

JP, 2010-56827, A

However, even when adjusting the resolution of an image using the estimated character size, a problem may occur when a character image is represented using new output image data.

The present specification discloses a technique capable of suppressing a defect when a character image is represented using output image data.

This specification discloses the following application examples, for example.

[Application example 1]
An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution larger than the resolution of interest to the reception unit,
And an output device.

According to this configuration, when outputting the output image data representing the character image, it is possible to reduce the data size of the output image data and to prevent the character from becoming unreadable due to too low resolution.
[Application example 2]
The output device according to Application Example 1,
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
The specific resolution is a resolution equal to or lower than the resolution of the original image data and higher than the resolution of interest.
Output device.
[Application example 3]
The output device according to application example 1 or 2, wherein
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a lower resolution, and generating the target image data with the updated target resolution,
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data using the attention image data corresponding to the lowest attention resolution among the attention resolutions satisfying the conditions.
Output device.

[Application example 4 ]
An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution lower than the target resolution to the reception unit,
And an output device.

According to this configuration, when outputting the output image data representing the character image, it is possible to make the character easy to read, and it is possible to prevent the data size of the output image data from becoming excessively large due to too high resolution. ..
[Application example 5]
The output device according to Application Example 4,
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
The specific resolution is a resolution equal to or higher than the resolution of the original image data and smaller than the resolution of interest.
Output device.
[Application example 6]
The output device according to Application Example 4 or 5,
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a higher resolution, and generating the target image data with the updated target resolution;
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data using the attention image data corresponding to the highest attention resolution among the attention resolutions satisfying the conditions.
Output device.
[Application example 7]
The output device according to application example 2 or 5, wherein
The resolution of interest has been updated at least once,
The specific resolution is the newest resolution of interest before the update,
Output device.
[Application example 8]
The output device according to application example 2 or 5, wherein
The output device in which the specific resolution is the same as the resolution of the original image data.
[Application example 9]
The output device according to any one of Application Examples 1 to 8,
The acquisition unit is a character image of interest that is the character image obtained using the image data of interest, and character recognition of the character image of interest having the same size as the size of the character image when the output image data is expanded. Is obtained as the attention reliability,
Output device.
[Application Example 10]
The output device according to any one of Application Examples 1 to 9,
The size of the character image when expanding the output image data is the same as the size of the character image when expanding the original image data,
Output device.
[Application example 11]
The output device according to any one of Application Examples 1 to 10,
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Output device.
[Application Example 12]
The output device according to any one of Application Examples 1 to 11,
The acquisition unit is a target character image that is the character image obtained by using the target image data, and has the same gradation number as the gradation number of the character image represented by the output image data. The reliability of character recognition of the target character image is acquired as the target reliability.
Output device.
[Application example 13]
The output device according to any one of Application Examples 1 to 12,
The gradation number of the character image represented by the output image data is smaller than the gradation number of the character image represented by the original image data,
Output device.
[Application Example 14]
The output device according to application example 2 or 5, wherein
The output control unit changes the allowable change amount according to the resolution of interest.
Output device.
[Application example 15]
The output device according to Application Example 3 or 6,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing a feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the initial value of the resolution of interest is determined using the resolution associated with the similar history information,
Output device.
[Application example 16]
The output device according to application example 15,
The original image data represents a character image of a plurality of pages,
The history management unit associates the resolution of the image data of interest used to output the output image data representing the character image of the first page with the feature information representing the feature of the character of the character image of the first page with a specific history. Storing information in the storage device,
When the characteristic information associated with the specific history information has the predetermined similar relationship with the characteristic of the character represented by the character image of the second page, the generation unit may generate the character image of the second page. The initial value of the resolution of interest of the image data of interest is determined using the resolution associated with the specific history information,
Output device.
[Application example 17]
The output device according to application example 15,
The original image data represents a character image of a plurality of pages,
The history management unit associates the resolution of the image data of interest used to output the output image data representing the character image of the first page with the feature information representing the feature of the character of the character image of the first page with a specific history. Storing information in the storage device,
The generation unit determines an initial value of a target resolution of target image data representing the character image of the second page by using a resolution associated with the specific history information.
Output device.
[Application Example 18]
The output device according to any one of Application Examples 1 to 17,
The generation unit generates a plurality of target image data corresponding to a plurality of different target resolutions,
The output control unit, the output image data including character data representing a result of character recognition of the character image obtained by using the attention image data corresponding to the highest attention resolution of the plurality of attention image data, Output,
Output device.
[Application example 19]
The output device according to any one of Application Examples 1 to 18,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing the feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the target resolution has an operation mode for determining using the resolution associated with the similar history information,
The output control unit has an operation mode of outputting the output image data by using the attention image data of the attention resolution regardless of whether or not the condition is satisfied.
Output device.

It should be noted that the technology disclosed in this specification can be realized in various modes, for example, an image data output method and an output device, a computer program for realizing the functions of those methods or devices, and It can be realized in the form of a recording medium (for example, a non-transitory recording medium) recording a computer program.

It is explanatory drawing which shows the data processing apparatus of an Example. It is a schematic diagram showing an example of output processing which outputs highly compressed PDF data. It is a flow chart which shows an example of data output processing. 7 is a flowchart of an example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a flowchart of another example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a flowchart of another example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a flowchart of another example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a flowchart of another example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a flowchart of another example of a binary image data generation process. It is a graph which shows the relationship between resolution RS and reliability RL. 7 is a graph showing a relationship between an allowable change amount and a resolution of interest. 9 is a part of a flowchart of another embodiment of a process of generating binary image data.

A. First embodiment:
FIG. 1 is an explanatory diagram illustrating a data processing device according to an embodiment. The data processing device 200 is, for example, a personal computer. The data processing device 200 includes a processor 210, a volatile storage device 220, a non-volatile storage device 230, a display unit 240 that displays an image, an operation unit 250 that receives a user operation, and a communication interface 270. is doing. These elements are connected to each other via a bus.

The processor 210 is a device that performs data processing, and is, for example, a CPU. The volatile storage device 220 is, for example, a DRAM, and the non-volatile storage device 230 is, for example, a flash memory.

The non-volatile storage device 230 stores a program 232 and history information 234. The processor 210 implements various functions by executing the program 232 (details will be described later). The processor 210 temporarily stores various intermediate data used for executing the program 232 in a storage device (for example, either the volatile storage device 220 or the non-volatile storage device 230). In this embodiment, the program 232 is stored in the non-volatile storage device 230 by the user. Details of the history information 234 will be described later.

The display unit 240 is a device that displays an image, and is, for example, a liquid crystal display. The operation unit 250 is a device that receives an operation by the user, and is, for example, a touch panel that is arranged on the display unit 240 in an overlapping manner. The user can input various instructions to the data processing device 200 by operating the operation unit 250.

The communication interface 270 is an interface for communicating with another device (for example, a USB interface, a wired LAN interface, an IEEE802.11 wireless interface).

The multifunction peripheral 300 is connected to the communication interface 270. The multi-function device 300 includes a communication interface 370, a scanner unit 380, a print execution unit 390, and a control unit 310 that controls these elements.

The communication interface 370 is an interface for communicating with another device. Here, the communication interface 370 of the multifunction peripheral 300 can communicate with the communication interface 270 of the data processing device 200. The scanner unit 380 optically scans an object such as a document using a photoelectric conversion element such as a CCD or a CMOS to generate scan data representing a read image (referred to as a “scan image”). The scan data is, for example, RGB bitmap data representing a color scan image. The color value of each color component of RGB is represented by M gradations from 0 to M−1 (M is an integer of 3 or more. For example, M=256). Hereinafter, image data of M gradations (M≧3) is also referred to as multi-valued data. The print execution unit 390 is a device that prints an image on a sheet (an example of a print medium) by a predetermined method (for example, a laser method or an inkjet method). The control unit 310 is, for example, a computer including a processor and a storage device. A program as firmware is stored in the storage device of the control unit 310 by the manufacturer of the multi-function peripheral 300. The processor of the controller 310 controls the elements 370, 380, 390 according to this firmware.

The data processing device 200 (specifically, the processor 210) can cause the print execution unit 390 of the multifunction peripheral 300 to print an image by communicating with the multifunction peripheral 300 according to a user's instruction. The scanner unit 380 of 300 can generate scan data. In the present embodiment, the data processing device 200 uses the scan data to generate the document data representing the object. Any format can be adopted as the format of the document data. Hereinafter, it is assumed that PDF (Portable Document Format) format document data is generated. In addition, in the present embodiment, the data processing device 200 generates PDF data with a reduced data size by compressing scan data (also referred to as “highly compressed PDF data”).

FIG. 2 is a schematic diagram showing an example of output processing for outputting highly compressed PDF data. In this embodiment, the image represented by the scan data i1 is a character image representing a character (also referred to as “original character image” or “original image”). Further, in the example of FIG. 2, the resolution of the scan data i1 is 300 dpi (dots per inch). In the figure, three types of highly compressed PDF data PDa, PDb, PDc that can be generated using the scan data i1 are shown.

The first type PDF data PDa is generated as follows. The processor 210 increases the resolution of the scan data i1 to generate high resolution data i2a, and binarizes the generated high resolution data i2a to generate high resolution binary data i3a. In the example of FIG. 2, the resolution of each of the high resolution data i2a and the high resolution binary data i3a is 600 dpi.

A known process can be adopted as the process of converting the resolution. For example, the color value of each color component at each pixel position after resolution conversion is calculated by interpolation. Further, in the binarization process, image data in which the total number of color components is 1 and the number of gradations is 2 is generated. A well-known process can be adopted as the binarization process. For example, the brightness value is calculated from the color value of each of the plurality of color components, and the binarized color value is determined by discriminating the brightness value with a predetermined threshold value.

The first type PDF data PDa includes high-resolution binary data i3a and a reduction command. The reduction command is a drawing command for reducing the size of the output image when the image is output (for example, displayed or printed) using the high resolution binary data i3a. An output image i4a in the figure represents an image that is output when the image is output using the first type PDF data PDa. The reduction command is configured such that the size of the output image i4a is the same as the size (referred to as “original size”) when the original image represented by the scan data i1 is output as it is. In the example of FIG. 2, the reduction command is, for example, a command that reduces the size (length) in half. The background color of the binary data is set to white and the character color is set to black. However, the processor 210 analyzes the scan data i1 to specify the background color and the character color, and the background and the character of the binary data represent the specified colors so that the first type PDF is displayed. The data PDa may be configured. The same applies to the other PDF data PDb and PDc.

The third type PDF data PDc is generated as follows. The processor 210 reduces the resolution of the scan data i1 to generate low resolution data i2c, and binarizes the generated low resolution data i2c to generate low resolution binary data i3c. In the example of FIG. 2, the resolution of each of the low resolution data i2c and the low resolution binary data i3c is 150 dpi. The third type PDF data PDc includes low-resolution binary data i3c and an enlargement command. The enlargement command is a drawing command for enlarging the size of the output image when the image is output using the low-resolution binary data i3c. An output image i4c in the drawing represents an image that is output when the image is output using the third type PDF data PDc. The enlargement command is configured such that the size of the output image i4c is the same as the original size. In the example of FIG. 2, the enlargement command is, for example, a command that doubles the size (length).

The second type PDF data PDb is generated as follows. The processor 210 binarizes the scan data i1 without changing the resolution to generate standard binary data i3b. The second type PDF data PDb includes standard binary data i3b. The second type PDF data PDb does not include the reduction command and the enlargement command. An output image i4b in the figure represents an image that is output when the image is output using the second type PDF data PDb. Since the resolution of the standard binary data i3b is the same as the resolution of the scan data i1, the size of the output image i4b is the same as the original size.

For all three types of PDF data PDa, PDb, PDc, a character image similar to the original image can be output in the same size as the original size. Further, in any of the three types of PDF data PDa, PDb, PDc, an image is represented using binary data i3a, i3b, i3c. Therefore, the data size can be significantly reduced as compared with the case where the scan data i1 of M gradations of the three color components (RGB) is used. When the resolution of the binary data is low, the character represented by the binary data is easily crushed, but the data size is small. On the other hand, when the resolution of the binary data is high, the data size is large, but the characters are hard to be crushed. As described above, the output process using the high resolution binary data and the output process using the low resolution binary data respectively have advantages. It is preferable that the data processing device 200 be equipped with at least one of an output process using high-resolution binary data and an output process using low-resolution binary data. In any output process, the resolution of binary data that can appropriately represent a character may differ depending on the scan data. In this embodiment, the data processing device 200 adjusts the resolution according to the scan data.

FIG. 3 is a flowchart showing an example of the data output process. The data output process is started when the user inputs a scan instruction to the operation unit 250 (FIG. 1) of the data processing device 200. In S100, the processor 210 acquires scan data. Specifically, the processor 210 transmits a scan instruction to the multi-function peripheral 300. The control unit 310 of the multi-function peripheral 300 drives the scanner unit 380 in response to the instruction to generate scan data. The control unit 310 transmits the generated scan data to the data processing device 200. The processor 210 of the data processing device 200 acquires the scan data from the multifunction peripheral 300. Hereinafter, it is assumed that the object read by the scanner unit 380 is a document representing a character. When the document includes N pages (N is an integer of 1 or more) of paper, the scan data represents N pages of character images.

In S105, the processor 210 acquires a portion representing a character image of one unprocessed page from the scan data (also referred to as page data). In S110, the processor 210 uses the page data to generate binary image data. Details of generation of the binary image data will be described later.

In S115, the processor 210 determines whether or not the processing of all pages has been completed. When the unprocessed page remains (S115: No), the processor 210 moves to S105 and executes the processing of the unprocessed page. Note that S117 in the figure is a step executed in another embodiment described later.

When the processing of all the pages is completed (S115: Yes), the processor 210 generates high compression PDF data using the binary data generated in S110 in S120. A drawing command may be added to the high compression PDF data according to the resolution of the binary data. Specifically, when the resolution of the binary data is lower than the resolution of the scan data, the enlargement command is added. When the resolution of the binary data is higher than the resolution of the scan data, the reduction command is added. When a character image is output (for example, displayed or printed) using highly compressed PDF data, the binary data is expanded according to a drawing command such as an enlargement command or a reduction command (for example, the resolution is converted). ). Further, as will be described later, in S110, character data representing the result of character recognition is also generated. The processor 210 embeds character data in the high compression PDF data. However, the character data may be omitted from the high compression PDF data.

In S130, the processor 210 outputs the generated high compression PDF data. Then, the processing of FIG. 3 ends. In the present embodiment, the processor 210 outputs the highly compressed PDF data to the nonvolatile storage device 230 (specifically, stores the highly compressed PDF data in the nonvolatile storage device 230). The output destination of the highly compressed PDF data may be any other storage device instead of the non-volatile storage device 230. For example, a storage device (for example, a network attached storage) connected to the communication interface 270 may be adopted. In addition, instead of storing the data in the storage device, the output causes the display device to display the image, causes the printing device to print the image, and transmits the data to an external device different from the data processing device 200. It may be done.

In general, the device to which the image data (here, highly compressed PDF data) is output may be any device that receives the data. As a device that receives data (also referred to as a receiving unit), for example, any one of a display device, a printing device, a data storage device, a communication interface for communicating with other devices, and various external devices may be adopted. .. Such a reception unit may be incorporated in an output device (here, the data processing device 200) that outputs image data, or may be an external device different from the output device instead. Further, outputting the image data to the reception unit may include processing associated with the function of the reception unit. The processing associated with the function of the reception unit is, for example, processing of displaying an image on a display device, processing of printing an image on a printing device, processing of storing image data in a data storage device, transmission of image data to a communication interface. Processing, etc.

FIG. 4 is a flowchart of an example of a binary image data generation process. FIG. 5 is a graph showing the relationship between the resolution RS and the reliability RL. The resolution RS on the horizontal axis indicates the resolution of the binary image data. The reliability RL on the vertical axis indicates the reliability of character recognition of a character image obtained using binary image data. High reliability indicates that the characters are easy to read and there is a low possibility of misrecognition. Generally, the higher the resolution RS, the harder it is for characters to be crushed. Therefore, the higher the resolution RS, the higher the reliability RL.

As a method of specifying the reliability RL, any known method can be adopted. For example, the character recognized by the character recognition process is rendered at the same resolution as the binary image data by referring to the font data representing the shape of the character. The first bitmap obtained by this rendering is superimposed on the corresponding character portion (called the second bitmap) of the bitmap represented by the binary image data. As a result, a part of the character pixel representing the character of the second bitmap overlaps the character pixel representing the character of the first bitmap (referred to as a superposition pixel). Here, the ratio of the total number of superimposed pixels to the total number of character pixels representing the characters of the second bitmap can be used as the reliability RL. When the character image represents multiple characters, the representative value (eg, average value, mode value, maximum value, minimum value, or median value) obtained using the reliability of each of the multiple characters is used. , May be used as the reliability RL.

As will be described later, in this embodiment, the resolution RS in the graph of FIG. 5 indicates the resolution in the binarization process. Then, the resolution of the binarized character image is converted to the original resolution RS1, and the reliability RL is specified by the character recognition process on the converted character image. In this way, the resolution in the character recognition processing is fixed to the original resolution RS1. Also in this case, the higher the resolution RS in the binarization process, the less the characters are crushed by the binarization process, and the higher the reliability RL.

In the process of FIG. 4, the resolution RS is updated from the resolution RS1 of the scan data i1 to a lower resolution (hereinafter, the resolution RS1 of the scan data i1 is also referred to as the original resolution RS1). The process of updating the resolution RS to a lower resolution is repeated until the reliability RL becomes less than the predetermined reference reliability Th. In the example of FIG. 5, the resolution RS is updated in the order of RS1, RSi1, RSi2, RSi3. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. The reliability RLb3 corresponding to the final resolution RSi3 is less than the reference reliability Th. Then, the lowest resolution among the resolutions whose reliability RL is equal to or higher than the reference reliability Th is selected as the final resolution (resolution RSi2 in the example of FIG. 5).

The details of FIG. 4 will be described below. In S200 of FIG. 4, the processor 210 binarizes the scan data i1 to generate binary data ia. The resolution of the binary data ia is the same as the resolution RS1 of the scan data i1.

In S210, the processor 210 executes character recognition processing of the binary data ia, and acquires the character data Ta representing the recognized character and the characteristic information Ca representing the characteristic of the recognized character. The character data Ta includes, for example, character code data representing a character (for example, Unicode data) and position data representing the position of the character. The characteristic information Ca represents, for example, a language, a character size, and a character font. Various known processes can be adopted as the character recognition process. A process called optical character recognition may be adopted.

In S220, the processor 210 determines whether the characteristic information Ca is similar to the learned characteristic. As will be described later, the history information 234 (FIG. 1) represents the correspondence between the feature information actually used for generating the binary image data and the resolution. The processor 210 refers to the history information 234 and searches for characteristic information having a predetermined similar relationship with the characteristic information Ca. As the similarity relationship, an arbitrary relationship indicating that the two pieces of feature information are similar can be adopted. For example, it may be adopted that the three attributes of language, character size, and character font are similar or identical. Further, it may be adopted that at least two attributes out of the three attributes of language, character size, and character font are similar or identical. In general, the fact that Q attributes (P is an integer of 1 or more) and Q attributes (Q is 1 or more and P or less) are similar or identical may be adopted as the similarity relationship. Here, the attributes being similar to each other means that the attribute difference between the two pieces of feature information is within the allowable range. For example, even if two fonts do not match, if each of the two fonts has a common design type (for example, a sans-serif type), the two fonts are similar. You can judge that If the difference between the two character sizes is within a predetermined range (for example, 2 points or less), it may be determined that the character sizes are similar. The allowable range is determined for each attribute. Further, the allowable range may be determined in advance or may be determined by the user.

When similar feature information having a similar relationship to the feature information Ca is detected from the history information 234 (S220: Yes), the processor 210 corresponds the attention resolution RSi to the similar feature information by the history information 234 in S230. It is set to the attached learned resolution RSp. Further, the processor 210 sets the candidate resolution RSs to the resolution RS1 of the scan data i1. Then, the processor 210 moves to S250.

If similar feature information is not detected from the history information 234 (S220: No), the processor 210 sets the resolution of interest RSi to the resolution obtained by subtracting a predetermined change amount dL from the resolution RS1 of the scan data i1 in S240. Set. Further, the processor 210 sets the candidate resolution RSs to the resolution RS1 of the scan data i1. Then, the processor 210 moves to S250.

In S250, the processor 210 sets the candidate data ic to the binary data ia of the resolution RS1. In S260, the processor 210 performs resolution conversion of the scan data i1 and generates multi-valued data ib1 of the target resolution RSi. In S270, the processor 210 binarizes the multi-valued data ib1 to generate the target binary data ib2. In S280, the processor 210 generates the binary data ib3 of the original resolution RS1 by enlarging the binary data of interest ib2. As the enlargement processing, resolution conversion processing for increasing the resolution is executed. In S290, the processor 210 executes the character recognition process of the binary data ib3 to specify the reliability RLb. The resolution RSi1 and the reliability RLb1 in FIG. 5 are examples of the resolution RSi of interest and the reliability RLb.

In S300, the processor 210 determines whether the reliability RLb is equal to or higher than the reference reliability Th. Hereinafter, the condition that the reliability RLb is equal to or higher than the reference reliability Th is also referred to as a first reliability condition. When RLb≧Th (S300: Yes), the processor 210 updates the resolution of interest RSi to a lower resolution in S310. Specifically, the updated resolution of interest RSi is the resolution obtained by subtracting the change amount dL from the updated resolution of interest RSi. Further, the processor 210 sets the candidate data ic to the binary value data ib2 of interest and sets the candidate resolution RSs to the resolution of interest RSi before updating. Then, the processor 210 moves to S260 and executes the processing of the remarked resolution RSi.

When the reliability RLb is less than the reference reliability Th (S300: NO), the processor 210 stops repeating the update of the resolution of interest RSi and moves to S320. At this stage, the candidate resolution RSs is set to the previous resolution of interest RSi, and the candidate data ic is set to the previous binary data of interest ib2. This candidate resolution RSs is the lowest resolution among the resolutions of interest RSi that satisfy the first reliability condition.

In the example of FIG. 5, when the resolution of interest RSi is RSi3, the reliability RLb3 is less than the reference reliability Th, and the update of the resolution of interest RSi stops. The candidate resolution RSs is set to the immediately preceding resolution RSi2, and the candidate data ic is set to the target binary data ib2 corresponding to the candidate resolution RSs (here, the resolution RSi2).

In S320 of FIG. 4, the processor 210 stores data representing the correspondence between the characteristic information Ca and the candidate resolutions RSs in the non-volatile storage device 230 as a part of the history information 234. In S330, the processor 210 outputs the candidate data ic to the storage device (for example, the volatile storage device 220 or the non-volatile storage device 230) as the finally generated binary data. Further, the processor 210 outputs the character data Ta (S210) as information indicating the character recognition result. Then, the process of FIG. 4 ends. If the determination result in S300 is No (for example, in S330), the processor 210 may generate image data of the candidate resolution RSs again and output the generated image data.

As described above, the present embodiment has the configuration and advantages described below.
(1-1) Highly compressed PDF data is output using the target binary data ib2 of the target resolution RSi that satisfies the first reliability condition that the reliability RLb is equal to or higher than the reference reliability Th. Then, when the resolution of interest RSi does not satisfy the first reliability condition, the high compression PDF data using the binary data of interest ib2 of the resolution of interest RSi is not output. Therefore, it is possible to reduce the data size of the high compression PDF data and to prevent the characters from being unreadable due to too low resolution.

(1-2) The processor 210 updates the resolution of interest RSi to a lower resolution (S310), and binarizes the updated resolution of interest RSi to generate binary data of interest ib2 (S260, S270). Obtaining the reliability RLb using the binary data ib2 (S280, S290) is repeated until the reliability RLb corresponding to the updated target resolution RSi does not satisfy the first reliability condition. Then, the processor 210 outputs the high compression PDF data by using the binary data of interest ib2 corresponding to the lowest resolution of the resolutions of interest RSi satisfying the first reliability condition. Therefore, the data size of the high compression PDF data can be appropriately reduced.

(1-3) As shown in S280 and S290 of FIG. 4, the processor 210 performs character recognition processing on the binary data ib3 obtained by converting the resolution of the binary data of interest ib2 into the original resolution RS1. , The reliability RLb corresponding to the target resolution RSi is acquired. In this case as well, the higher the resolution when performing binarization, the less the characters are crushed due to binarization, and the higher the reliability RLb. Further, as described in S120 of FIG. 3, when a character image is displayed or printed using the high compression PDF data, the binary data included in the high compression PDF data is expanded according to the drawing command. Here, to develop image data means to determine (for example, calculate, specify) data representing the color value of each of a plurality of pixels used for display or printing for display or printing using the image data. It means that. By developing the image data, for example, data representing the color values of each of the plurality of pixels is arranged in the memory. The size of the character image represented by the expanded data is the same as the size of the original image represented by the scan data i1 having the original resolution RS1. Here, when two images representing the same object have the same resolution, the two images have the same size. As described above, the character image to be subjected to the character recognition process is a character image obtained by using the binary data ib2 of interest, and is a character image having the same size as the character image when the high compression PDF data is expanded. .. Therefore, it is possible to use the reliability RLb suitable for the size of the character image when the highly compressed PDF data is expanded. As a result, it is possible to prevent the characters represented by using the high compression PDF data from becoming difficult to read.

(1-4) The size of the character image at the time of expanding the high compression PDF data is the same as the size of the character image when displaying or printing the character image using the scan data i1. Therefore, the high compression PDF data can represent a character image having the same size as the scan data i1. Here, the high compression PDF data includes a command for enlarging the character image represented by the binary data. Therefore, the high compression PDF data can represent a character image of an appropriate size.

(1-5) As described in S290, the gradation number of the character image (image data ib3) that is the target of the character recognition process for acquiring the reliability RLb is the same as that of the character image represented by the high compression PDF data. It is the same as the number of gradations (here, binary). Therefore, the reliability RLb suitable for the number of gradations of the character image represented by the high compression PDF data can be used.

(1-6) The gradation number of the character image represented by the scan data i1 is M (M≧3), and the gradation number of the character image represented by the high compression PDF data is 2. In this way, since the number of gray levels of the high compression PDF data is smaller than the number of gray levels of the scan data i1, the data size of the high compression PDF data can be appropriately reduced.

(1-7) As described in S320 of FIG. 4, the processor 210 stores the history information 234 indicating the correspondence relationship between the resolution satisfying the first reliability condition and the characteristic information Ca in the nonvolatile storage device 230. .. Then, as described in S220 and S230, the processor 210 stores similar history information, which is history information associated with similar feature information having a predetermined similarity relationship with the feature information Ca, in the non-volatile storage device 230. If there is, the initial value of the resolution of interest RSi is determined as the resolution associated with the similar history information. Therefore, the processor 210 can use a resolution suitable for the character features. The initial value of the resolution of interest RSi is the first resolution of interest RSi that is the target of the determination of whether or not to continue updating the resolution of interest RSi (for example, the determination based on the reliability condition of S300 in FIG. 4), For example, it is the first resolution when the target resolution RSi is set to a resolution different from the original resolution RS1. Since the original resolution RS1 can also be used for outputting the high compression PDF data, the original resolution RS1 is also one of the resolutions of interest.

(1-8) As described in FIG. 3, when the scan data represents a plurality of pages of character images, the process of S110 is repeated for each page. Therefore, in the process of the first page, the specific history information indicating the correspondence between the resolution and the feature information is added to the history information 234. In the process of the second page, when the character feature is similar to the feature associated with the specific history information, the initial value of the resolution of interest RSi for the second page is the resolution of the specific history information. It is decided using. Therefore, it is possible to reduce the load of the process of outputting the output image data by using the scan data i1 representing the character images of a plurality of pages.

(1-9) As shown in S210 and S260 to S290 of FIG. 4, the processor 210 generates binary data ia and ib2 corresponding to a plurality of target resolutions including the original resolution RS1. Then, as described in S330 (FIG. 4) and S120 (FIG. 3), the processor 210 performs the processing of the character image obtained using the binary data ia corresponding to the highest resolution RS1 of the plurality of binary data. Highly compressed PDF data including character data Ta representing the result of character recognition is output. Therefore, the high compression PDF data can include the appropriate character data Ta.

B. Second embodiment:
FIG. 6 is a flowchart of another example of the binary image data generation process. FIG. 7 is a graph showing the relationship between the resolution RS and the reliability RL. Similar to FIG. 5, FIG. 7 shows changes in the resolution of interest RSi and the reliability RLb. The only difference from the embodiments of FIGS. 4 and 5 is that S210 and S300 are replaced by S210a and S300a. The other part of the process of FIG. 6 is the same as the process of FIG. Of the steps in FIG. 6, the same steps as those in FIG. 4 are designated by the same reference numerals, and the description thereof will be omitted.

In S210a, the processor 210 executes a character recognition process to acquire the reliability RLa in addition to the character data Ta and the characteristic information Ca (also referred to as the original reliability RLa). In S300a, the processor 210 uses the reference reliability Tha represented by using the original reliability RLa instead of the predetermined reference reliability Th. In this embodiment, the reference reliability Tha is the reliability obtained by subtracting the allowable change amount Rta from the original reliability RLa. The allowable change amount Rta is predetermined. Hereinafter, the condition that the reliability RLb is equal to or higher than the reference reliability Tha is also referred to as a second reliability condition. The process of updating the resolution of interest RSi to a lower resolution is repeated until the second reliability condition is no longer satisfied.

In the example of FIG. 7, the resolution RS is updated in the order of RS1, RSi1, RSi2, and RSi3, as in the example of FIG. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. The reliability RLb3 corresponding to the final resolution RSi3 is less than the reference reliability Tha. Then, the lowest resolution among the resolutions whose reliability RL is equal to or higher than the reference reliability Tha is selected as the final resolution (resolution RSi2 in the example of FIG. 7).

In this way, the original reliability RLa obtained by using the binary data ia having the same original resolution RS1 as the resolution of the scan data i1 is used to determine the reference reliability Tha to be compared with the reliability RLb. Therefore, it is possible to output highly compressed PDF data including binary data having a resolution suitable for the scan data i1.

The process of the second embodiment is the same as the process of the first embodiment except that the reference reliability Tha is used. Therefore, the second embodiment has the same various advantages as the first embodiment (for example, the advantages described in (1-1) to (1-9)).

C. Third embodiment:
FIG. 8 is a flowchart of another example of the binary image data generation process. FIG. 9 is a graph showing the relationship between the resolution RS and the reliability RL. Similar to FIG. 5, FIG. 9 shows changes in the resolution of interest RSi and the reliability RLb. The only difference from the embodiments of FIGS. 6 and 7 is that S250, S300a and S310 are replaced with S250b, S300b and S310b. The other part of the process of FIG. 8 is the same as the process of FIG. Of the steps in FIG. 8, the same steps as those in FIG. 6 are designated by the same reference numerals, and the description thereof will be omitted.

In S250b, the processor 210 sets the previous reliability RLc to the original reliability RLa in addition to the process of S250 of FIG. The previous reliability level RLc is used to indicate the reliability level of the resolution of interest immediately preceding the current resolution of interest RSi. That is, when the resolution of interest RSi is updated one or more times, the previous reliability RLc is the latest resolution of interest RSi of the resolutions of interest RSi before the last update.

In S300b, the processor 210 uses the reliability for judgment Thb represented by using the previous reliability RLc instead of the reference reliability Tha. In the present embodiment, the determination reliability Thb is the reliability obtained by subtracting the allowable change amount Rtb from the previous reliability RLc. The allowable change amount Rtb is predetermined. Hereinafter, the condition that the reliability RLb is equal to or higher than the determination reliability Thb is also referred to as a third reliability condition.

In S310b, the processor 210 updates the previous reliability RLc to the reliability RLb in addition to the process of S310 of FIG. Accordingly, in S260-S300b for the updated resolution of interest RSi, the previous reliability RLc represents the reliability RLb corresponding to the previous (that is, before the update) resolution of interest RSi.

The determination reliability Thb is updated every time the target resolution RSi is updated. In the example of FIG. 9, the resolution RS is updated in the order of RS1, RSi1, RSi2, RSi3. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. Here, the determination reliability Thb1 that is compared with the reliability RLb1 is the original reliability RLa−the allowable change amount Rtb. The determination reliability Thb2 to be compared with the next reliability RLb2 is reliability RLb1−allowable change amount Rtb. The determination reliability Thb3 to be compared with the next reliability RLb3 is reliability RLb2-allowable change amount Rtb. The reliability RLb3 corresponding to the final resolution RSi3 is less than the determination reliability Thb3. The lowest resolution among the resolutions whose reliability RL is equal to or higher than the determination reliability Thb is selected as the final resolution (resolution RSi2 in the example of FIG. 9).

As described above, when the resolution of attention RSi is reduced, the reliability RLb is reduced. Then, as the reliability RLb decreases, the determination result of S300b in FIG. 8 changes (here, “Yes” to “No”), and the resolution finally used is determined (S330). From a different point of view, such processing can be explained as follows. As shown in FIG. 9, the rate of change in the reliability RL with respect to the change in the resolution RS (that is, the slope) is larger as the resolution RS is smaller. Therefore, the reduction amount of the reliability RLb when the target resolution RSi is reduced by the change amount dL is larger as the target resolution RSi is smaller. Here, the resolution RS in which the reduction amount of the reliability RL when the resolution RS is reduced by the change amount dL is the same as the allowable change amount Rtb is referred to as a first resolution RSx. The reliability Thx corresponding to the first resolution RSx is called the first reliability Thx. In this case, when the resolution RS is reduced from “RSx+dL” to “RSx”, the reliability RL is reduced from “Thx+Rtb” to “Thx”.

When the resolution of interest RSi is equal to or higher than the first resolution RSx, that is, when the reliability RLb is equal to or higher than the first reliability Thx, the reduction amount of the reliability RL when the resolution RS is reduced by the variation amount dL is the allowable variation amount. It is less than or equal to Rtb. Therefore, in S300b of FIG. 8, the reliability RLb is equal to or higher than the determination reliability Thb (S300b: Yes). On the other hand, when the resolution of interest RSi is less than the first resolution RSx, that is, when the reliability RLb is less than the first reliability Thx, the reduction amount of the reliability RL when the resolution RS is reduced by the change amount dL is allowable. The change amount is larger than Rtb. Therefore, in S300b of FIG. 8, the reliability RLb is less than the determination reliability Thb (S300b: No).

In this way, the condition for shifting from S300b to S310b and updating the resolution of interest RSi is that "the reliability RLb is equal to or higher than the first reliability Thx". As described above, the third reliability condition of S300b is equivalent to the condition that "the reliability RLb is equal to or higher than the first reliability Thx". Similarly, the condition for shifting from S300b to S320 is equivalent to the condition that "the reliability RLb is less than the first reliability Thx". Thus, it can be said that the first reliability Thx is the reference reliability indicating the reference of the reliability RLb.

The first reliability Thx is specified by two pairs of the resolution RS and the reliability RL, as described above. These two pairs are two pairs in which the difference in resolution RS is the change amount dL and the difference in reliability RL is the allowable change amount Rtb. Such two pairs are a pair of resolution “RSx+dL” and reliability “Thx+Rtb” and a pair of resolution “RSx” and reliability “Thx”. Here, as shown in FIG. 9, the larger resolution “RSx+dL” is equal to or lower than the original resolution RS1. Also, when viewed from the resolution of interest finally adopted (for example, the resolution RSi2 of FIG. 9), the larger resolution “RSx+dL” is larger than the resolution of interest RSi2. The first reliability Thx is a reliability obtained by subtracting the allowable change amount Rtb from the specific reliability “Thx+Rtb” corresponding to the specific resolution “RSx+dL”. Therefore, it can be said that the first reliability Thx is expressed using the specific reliability “Thx+Rtb” corresponding to the specific resolution “RSx+dL” and the allowable change amount Rtb.

Further, in this embodiment, the determination reliability Thb is a reliability obtained by subtracting the allowable change amount Rtb from the previous reliability RLc associated with the pre-update target resolution RSi. The resolution of interest RSi before the update is equal to or lower than the original resolution RS1 of the scan data i1 and larger than the current resolution of interest RSi. Since the current determination reliability Thb is determined using the pre-reliability RLc corresponding to the remarked resolution RSi before updating, it is possible to determine the resolution suitable for the characteristics of the character image at various resolutions. is there. As a result, appropriate highly compressed PDF data can be output.

The processing of the third embodiment is the same as the processing of the first and second embodiments, except that the determination reliability Thb is used. Therefore, the third embodiment has the same various advantages as the first and second embodiments (for example, the advantages described in (1-1) to (1-9)).

D. Fourth embodiment:
FIG. 10 is a flowchart of another example of the binary image data generation process. FIG. 11 is a graph showing the relationship between the resolution RS and the reliability RL. Similar to FIG. 5, FIG. 11 shows changes in the resolution of interest RSi and the reliability RLb. Unlike the embodiments of FIGS. 4 and 5, in the present embodiment, the resolution of interest RSi is updated from the original resolution RS1 to a higher resolution (S240c, S310c). Then, the process of updating the resolution of interest RSi to a higher resolution is repeated until the fourth reliability condition "the reliability RLb is less than the reference reliability Thc" is not satisfied (S300c). Here, the reference reliability Thc is a predetermined reliability.

Specifically, the difference from the embodiment of FIGS. 4 and 5 is that S240, S280, S290, S300, S310, S330 are replaced by S240c, S280c, S290c, S300c, S310c, S330c. is there. The other part of the process of FIG. 10 is the same as the process of FIG. Of the steps in FIG. 10, steps that are the same as the steps in FIG. 4 are given the same reference numerals, and descriptions thereof are omitted.

In S240c, the processor 210 sets the resolution of interest RSi to the resolution obtained by adding a predetermined change amount dH to the original resolution RS1. In S280c, the processor 210 reduces the binary data ib2 of interest to generate the binary data ib3 of the original resolution RS1. As the reduction processing, resolution conversion processing for lowering the resolution is executed. In S290c, the processor 210 executes character recognition processing of the binary data ib3, and specifies the reliability RLb and the character data Tb.

In S300c, the processor 210 determines whether or not the fourth reliability condition is satisfied. When the fourth reliability condition is satisfied (S300c: Yes), the processor 210 moves to S310c. In S310c, the processor 210 updates the resolution of interest RSi to a resolution obtained by adding the change amount dH to the resolution of interest RSi. The updating of the candidate data ic and the candidate resolution RSs is the same as the updating in S310 of FIG. After S310c, the processor 210 moves to S260 and executes the processing of the remarked resolution RSi.

When the reliability RLb is equal to or higher than the reference reliability Thc (S300cNo), the processor 210 stops repeating the update of the resolution of interest RSi and moves to S320. S320 is the same as S320 of FIG. In S330c, the processor 210 outputs the candidate data ic to the storage device (for example, the volatile storage device 220 or the non-volatile storage device 230) as the finally generated binary data. Further, the processor 210 outputs the character data Tb (S290c) as information indicating the character recognition result. Then, the processing of FIG. 10 ends.

In the example of FIG. 11, the resolution RS is updated in the order of RS1, RSi1, RSi2, RSi3. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. The reliability RLb3 corresponding to the final resolution RSi3 is greater than or equal to the reference reliability Thc. Then, the highest resolution among the resolutions whose reliability RL is lower than the reference reliability Th is selected as the final resolution (resolution RSi2 in the example of FIG. 11).

As described above, the present embodiment has the configuration and advantages described below.
(4-1) The high compression PDF is output using the target binary data ib2 of the target resolution RSi that satisfies the fourth reliability condition that the reliability RLb is less than the reference reliability Thc. Then, if the resolution of interest RSi does not satisfy the fourth reliability condition, the high compression PDF data using the binary data of interest ib2 of the resolution of interest RSi is not output. Therefore, when outputting the highly compressed PDF data, it is possible to prevent the resolution from becoming too high and the data size of the highly compressed PDF data becoming excessively large.

(4-2) The processor 210 updates the resolution of interest RSi to a higher resolution (S310c), and binarizes the updated resolution of interest RSi to generate binary data of interest ib2 (S260, S270). Obtaining the reliability RLb using the binary data ib2 (S280c, S290c) is repeated until the reliability RLb corresponding to the updated target resolution RSi does not satisfy the fourth reliability condition. Then, the processor 210 outputs the high compression PDF data using the binary data of interest ib2 corresponding to the highest resolution of attention RSi among the resolutions of attention RSi satisfying the fourth reliability condition. Therefore, it is possible to properly read the characters represented by the high compression PDF data.

(4-3) Highly compressed PDF data includes a command for reducing the character image represented by the binary data. Therefore, the high compression PDF data can represent a character image of an appropriate size.

(4-4) The character data Tb output in S330c is the character data Tb generated in the last S290c of the repeatable S290c. That is, the character data Tb is associated with the highest resolution RSi of the plurality of resolutions RS for which the binary data ib2 of interest has been generated. In the example of FIG. 11, the character data Tb is associated with the resolution RSi3. Therefore, the highly compressed PDF data can include the appropriate character data Tb.

The processing of the fourth embodiment is the same as the processing of the first embodiment except for the above differences. Therefore, the fourth embodiment has the same various advantages as the first embodiment (for example, the advantages described in (1-3) to (1-9)).

E. Fifth embodiment:
FIG. 12 is a flowchart of another example of the binary image data generation process. FIG. 13 is a graph showing the relationship between the resolution RS and the reliability RL. Similar to FIG. 11, FIG. 13 shows changes in the resolution of interest RSi and the reliability RLb. The only difference from the embodiments of FIGS. 10 and 11 is that S210 and S300c are replaced by S210a and S300d. The other part of the process of FIG. 12 is the same as the process of FIG. Of the steps in FIG. 12, the same steps as those in FIG. 10 are designated by the same reference numerals, and the description thereof will be omitted.

S210a is the same as S210a in FIG. 6 (the original reliability RLa is specified in addition to the character data Ta and the characteristic information Ca). In S300d, the processor 210 uses the reference reliability Thd represented by using the original reliability RLa instead of the predetermined reference reliability Thc. In this embodiment, the reference reliability Thd is the reliability obtained by adding the allowable change amount Rtd to the original reliability RLa. The allowable change amount Rtd is predetermined. Hereinafter, the condition that the reliability RLb is less than the reference reliability Thd is also referred to as a fifth reliability condition. The process of updating the resolution of interest RSi to a higher resolution is repeated until the fifth reliability condition is no longer satisfied.

In the example of FIG. 13, similarly to the example of FIG. 11, the resolution RS is updated in the order of RS1, RSi1, RSi2, RSi3. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. The reliability RLb3 corresponding to the final resolution RSi3 is larger than the reference reliability Thd. Then, the highest resolution among the resolutions whose reliability RL is less than the reference reliability Thd is selected as the final resolution (resolution RSi2 in the example of FIG. 13).

As described above, the original reliability RLa obtained by using the binary data ia having the same original resolution RS1 as the resolution of the scan data i1 is used to determine the reference reliability Thd to be compared with the reliability RLb. Therefore, it is possible to output highly compressed PDF data including binary data having a resolution suitable for the scan data i1.

The process of the fifth embodiment is the same as the process of the fourth embodiment except that the reference reliability Thd is used. Therefore, the fifth embodiment has the same various advantages as the fourth embodiment (for example, the advantages described in (1-3)-(1-9) and (4-1)-(4-4)). .

F. Sixth embodiment:
FIG. 14 is a flowchart of another example of the binary image data generation process. FIG. 15 is a graph showing the relationship between the resolution RS and the reliability RL. Similar to FIG. 13, FIG. 15 shows changes in the resolution of interest RSi and the reliability RLb. The only difference from the embodiments of FIGS. 12 and 13 is that S250, S300d and S310c are replaced with S250b, S300f and S310f. S250b is the same as S250b in FIG. The other part of the process of FIG. 14 is the same as the process of FIG. Of the steps in FIG. 14, the same steps as those in FIG. 12 are designated by the same reference numerals, and the description thereof will be omitted.

In S300f, the processor 210 uses the reliability for judgment Thf represented by using the previous reliability RLc instead of the reference reliability Thd. In the present embodiment, the determination reliability Thf is the reliability obtained by adding the allowable change amount Rtf to the previous reliability RLc. The allowable change amount Rtf is predetermined. The condition for the determination result of S300f to be Yes is “reliability RLb≧reliability for determination Thf”. Hereinafter, the condition that the reliability RLb is equal to or higher than the determination reliability Thf is also referred to as a sixth reliability condition.

In S310f, the processor 210 updates the previous reliability RLc to the reliability RLb in addition to the process of S310c in FIG. As a result, in S260-S300f for the updated resolution of interest RSi, the previous reliability RLc represents the reliability RLb corresponding to the previous (that is, before the update) resolution of interest RSi.

The determination reliability Thf is updated each time the target resolution RSi is updated. In the example of FIG. 15, the resolution RS is updated in the order of RS1, RSi1, RSi2, RSi3. The reliability levels RLa, RLb1, RLb2, and RLb3 correspond to the resolutions RS1, RSi1, RSi2, and RSi3, respectively. Here, the determination reliability Thf1 compared with the reliability RLb1 is the original reliability RLa+the allowable change amount Rtf. The determination reliability Thf2, which is compared with the next reliability RLb2, is the reliability RLb1+the allowable change amount Rtf. The determination reliability Thf3 to be compared with the next reliability RLb3 is the reliability RLb2+the allowable change amount Rtf. The reliability RLb3 corresponding to the final resolution RSi3 is less than the determination reliability Thf3. The highest resolution among the resolutions whose reliability RL is equal to or higher than the reference reliability Thf is selected as the final resolution (resolution RSi2 in the example of FIG. 15).

In this way, as the resolution of attention RSi increases, the reliability RLb also increases. Then, as the reliability RLb increases, the determination result of S300f in FIG. 14 changes (here, “Yes” to “No”), and the resolution finally used is determined (S330c). From a different point of view, such processing can be explained as follows. As shown in FIG. 15, the rate of change in reliability RL with respect to change in resolution RS (that is, the slope) is smaller as the resolution RS is larger. Therefore, the increase amount of the reliability RLb when the target resolution RSi is increased by the change amount dH is smaller as the target resolution RSi is larger. Here, the resolution RS at which the increase amount of the reliability RL when the resolution RS increases by the change amount dH is the same as the allowable change amount Rtf is referred to as the second resolution RSy. The reliability Thy corresponding to the second resolution RSy is called the second reliability Thy. In this case, when the resolution RS increases from "RSy-dH" to "RSy", the reliability RL increases from "Thy-Rtf" to "Thy".

When the resolution of interest RSi is less than or equal to the second resolution RSy, that is, when the reliability RLb is less than the second reliability Thy, the increase amount of the reliability RL when the resolution RS increases by the change amount dH is the allowable change amount. It is Rtf or more. Therefore, in S300f of FIG. 14, the reliability RLb is equal to or higher than the determination reliability Thf (S300f: Yes). On the other hand, when the resolution of interest RSi is larger than the second resolution RSy, that is, when the reliability RLb is larger than the second reliability Thy, the increase amount of the reliability RL when the resolution RS increases by the change amount dH is an allowable change. It is smaller than the amount Rtf. Therefore, in S300f of FIG. 14, the reliability RLb is less than the determination reliability Thf (S300f: No).

In this way, the condition for shifting from S300f to S310f and updating the resolution of interest RSi is the condition that “the reliability RLb is less than the second reliability Thy”. Thus, the sixth reliability condition of S300f is equivalent to the condition that "the reliability RLb is less than the second reliability Thy". Similarly, the condition for shifting from S300f to S320 is equivalent to the condition that “the reliability RLb is equal to or higher than the second reliability Thy”. In this way, it can be said that the second reliability Thy is the reference reliability indicating the reference of the reliability RLb.

The second reliability Thy is specified by two pairs of the resolution RS and the reliability RL, as described above. These two pairs are two pairs in which the difference in resolution RS is the change amount dH and the difference in reliability RL is the allowable change amount Rtf. Such two pairs are a pair of resolution “RSy-dH” and reliability “Thy-Rtf”, and a pair of resolution “RSy” and reliability “Thy”. Here, as shown in FIG. 15, the smaller resolution "RSy-dH" is the original resolution RS1 or more. Further, when viewed from the resolution of attention finally adopted (for example, the resolution RSi2 of FIG. 15), the smaller resolution “RSy-dH” is smaller than the resolution of attention RSi2. The second reliability Thy is a reliability obtained by adding the allowable change amount Rtf to the specific reliability “Thy-Rtf” corresponding to the specific resolution “RSy-dH”. Therefore, it can be said that the second reliability Thy is expressed using the specific reliability “Thy-Rtf” corresponding to the specific resolution “RSy-dH” and the allowable change amount Rtf.

Thus, in this embodiment, the processor 210 updates the resolution of interest RSi to a higher resolution (S310f), and generates the binary data of interest ib2 by binarization at the updated resolution of interest RSi (S260, (S270), the reliability RLb is acquired using the binary data of interest ib2 (S280c, S290c) is repeated until the reliability RLb corresponding to the updated resolution of interest RSi does not satisfy the sixth reliability condition. Then, the processor 210 uses the target binary data ib2 corresponding to the highest target resolution RSi of the target resolutions RSi satisfying the sixth reliability condition (in other words, the reliability RLb<the second reliability Thy). , Highly compressed PDF data is output. Therefore, it is possible to properly read the characters represented by the high compression PDF data.

The determination reliability Thf is a reliability obtained by adding the allowable change amount Rtf to the previous reliability RLc associated with the pre-update target resolution RSi. The resolution of interest RSi before update is equal to or higher than the original resolution RS1 of the scan data i1 and smaller than the current resolution of interest RSi. Since the current determination reliability Thf is determined using the pre-reliability RLc corresponding to the target resolution RSi before updating, it is possible to determine the resolution suitable for the characteristics of the character image at various resolutions. is there. As a result, appropriate highly compressed PDF data can be output.

The processing of the sixth embodiment is the same as the processing of the fifth embodiment except for the above differences. Therefore, the sixth embodiment has the same various advantages as the fifth embodiment (for example, the advantages described in (1-3)-(1-9) and (4-1)-(4-4)). .

G. Seventh embodiment:
FIG. 16A is a graph showing another example of the allowable change amounts Rta (FIG. 6) and Rtb (FIG. 8), and FIG. 16B is the graph showing the allowable change amounts Rtd (FIG. 12) and Rtf( 15 is a graph showing another example of (FIG. 14). In FIG. 16A, the horizontal axis indicates the resolution of interest RSi, and the vertical axis indicates the allowable change amounts Rta and Rtb. In FIG. 16B, the horizontal axis represents the resolution of interest RSi, and the vertical axis represents the allowable change amounts Rtd and Rtf. In the present embodiment, these allowable change amounts Rta, Rtb, Rtd, Rtf change according to the resolution RSi of interest.

In the embodiment shown in FIG. 16A, the allowable change amounts Rta and Rtb are smaller as the target resolution RSi is smaller (that is, the difference between the original resolution RS1 and the target resolution RSi is larger). Therefore, in the example of FIG. 7, the reference reliability Tha increases as the target resolution RSi decreases. In the example of FIG. 9, the smaller the target resolution RSi, the greater the determination reliability Thb with respect to the reliability RLb before updating. In other words, in the example of FIG. 9, the smaller the resolution of interest RSi, the larger the first reliability Thx. Therefore, it is possible to prevent the noticeable resolution RSi from becoming too small and making it difficult to read characters.

In the embodiment shown in FIG. 16B, the allowable change amounts Rtd and Rtf are smaller as the target resolution RSi is larger (that is, the difference between the original resolution RS1 and the target resolution RSi is larger). Therefore, in the example of FIG. 13, the reference reliability Thd decreases as the resolution of attention RSi increases. In the example of FIG. 15, the reference reliability Thf with respect to the reliability RLb before updating decreases as the resolution of attention RSi increases. In other words, in the example of FIG. 15, the second reliability Thy becomes smaller as the resolution of interest RSi is larger. Therefore, it is possible to prevent the resolution of attention RSi from becoming excessively large and the data size from becoming large.

Note that the correspondence relationship between the resolution of interest RSi and the allowable change amounts Rta, Rtb, Rtd, and Rtf can be replaced with the correspondence relationships shown in FIGS. 16A and 16B, and various other correspondence relationships can be adopted. Is. For example, the permissible change amounts Rta, Rtb, Rtd, and Rtf may change stepwise with respect to the change in the resolution of interest RSi.

H. Eighth Example:
In the eighth embodiment, S117 of FIG. 3 is effective. That is, if the determination result in S115 is No, in S117, the processor 210 fixes the initial value of the resolution of attention RSi on the second and subsequent pages to the resolution of attention RSi finally used in the processing of page 1. For example, when the processing of FIG. 4 is executed as S110, the processor 210 highly compresses the initial value of the resolution of interest RSi by the processing of one page in S230 and S240 regardless of the determination result of S220 of FIG. The resolution is set to be the same as the resolution of interest RSi used for outputting the PDF data. The process of S110 of FIG. 3 may be the process of the other embodiment described above. In any case, the initial value of the resolution of attention RSi on the second and subsequent pages is determined by using the resolution of attention RSi used for outputting the highly compressed PDF data of the first page.

Usually, the character images of a plurality of pages of one document represent characters having the same characteristics. Therefore, the attention resolution RSi suitable for the first page is often suitable for the second page and thereafter. Therefore, when the attention resolution RSi of the first page is used as the initial value, the frequency of updating the attention resolution RSi can be reduced. As a result, the processing load can be reduced when the scan data represents a plurality of pages of character images.

I. Ninth embodiment:
FIG. 17 is a part of a flowchart of another embodiment of the binary image data generation processing. The difference from the flowcharts of FIGS. 4, 6, 8, 10, 12, and 14 is only that the processing is changed to S430-S490 when the determination result of S220 is Yes. The process when the determination result of S220 is No is the same as the process of each of the above-described embodiments, and the process proceeds from S220 to S240 or S240c.

In S430, the processor 210 sets the noted resolution RSi to the learned resolution RSp, as in S230 of each of the above-described embodiments. S460 and S470 are the same as S260 and S270 of the above-described embodiments, respectively. By these S460 and S470, the processor 210 generates the target binary data ib2 of the target resolution RSi. In S490, the processor 210 outputs the binary data of interest ib2, as in S330 and S330c of the above-described embodiments. Note that in S490, the expanded or reduced binary data ib3 (that is, the binary data ib3 of the original resolution RS1) generated in S280 and S280c may be output.

As described above, the processor 210, when the similar history information, which is the history information associated with the similar feature information having a predetermined similarity relationship with the feature information Ca, is stored in the nonvolatile storage device 230, the resolution of interest. RSi is determined as the learned resolution RSp associated with the similar history information. When the feature information is similar, the appropriate resolution is also estimated to be approximately the same. Therefore, the processor 210 can output the high compression PDF data with the resolution suitable for the character by using the learned resolution RSp.

J. Modification:
(1) As the procedure of the binary image data generation processing, various other procedures can be adopted instead of the above procedure. For example, in S300b of FIG. 8, the processor 210 determines whether or not the difference obtained by subtracting the "reliability RLb-original reliability RLa" from the "previous reliability RLc-original reliability RLa" is larger than the allowable change amount Rtb. May be judged. Further, in S300f of FIG. 14, the processor 210 determines whether or not the difference obtained by subtracting the “previous reliability RLc-original reliability RLa” from the “reliability RLb-original reliability RLa” is less than the allowable change amount Rtf. May be judged. Also, the change amount dL in FIGS. 4, 6, and 8 and the change amount dH in FIGS. 10, 12, and 14 may be variable values instead of fixed values. For example, the change amounts dL and dH may be reduced as the number of updates of the resolution of interest RSi is increased.

Further, the reliability RLb corresponding to the resolution of interest RSi used in S300, S300a, S300b, S300c, S300d, and S300f of FIGS. 4, 6, 8, 10, 12, and 14 is the high compression PDF data. The reliability of character recognition of a character image having a size different from the size of the character image at the time of expansion may be used. For example, as the reliability RLb used in S300, S300a, and S300b of FIGS. 4, 6, and 8, the reliability of character recognition of a small character image represented by the target binary data ib2 of S270 may be adopted. .. Further, as the reliability RLb used in S300c, S300d, and S300f of FIGS. 10, 12, and 14, the reliability of character recognition of a large character image represented by the noted binary data ib2 of S270 may be adopted. ..

Further, it is determined whether the reliability RLb corresponding to the resolution of interest RSi satisfies the reliability condition (for example, S300, S300a, S300b, and S300c in FIGS. 4, 6, 8, 10, 12, and 300). , S300d, S300f) may be only once. For example, the processor 210 outputs the target binary data ib2 when the reliability RLb satisfies the reliability condition, and updates the target resolution RSi without updating when the reliability RLb does not satisfy the reliability condition. Other image data (also referred to as alternative data) may be output. Generally, the maximum value U of the number of determinations may be predetermined (U is an integer of 1 or more). In this case, the processor 210 updates the target resolution RSi, and determines whether or not the reliability RLb corresponding to the updated target resolution RSi satisfies the reliability condition, repeating a series of processes up to U times. .. The processor 210 may output the alternative data when the reliability condition is not satisfied despite the series of processes being repeated U times.

As alternative data that is output when the reliability condition is not satisfied, various data different from the target binary data ib2 of the target resolution RSi can be adopted. For example, image data having the same resolution as the original resolution RS1 (for example, binary data ia or scan data i1) may be output. When the resolution of interest RSi is lower than the original resolution RS1 as in the embodiments of FIGS. 4, 6 and 8, image data of a resolution higher than the resolution of interest RSi may be output. For example, image data having a resolution higher than the original resolution RS1 may be output, or image data having a resolution higher than the target resolution RSi and lower than the original resolution RS1 may be output. As an example, it is possible to output image data having a resolution that is an intermediate value between the resolution of interest RSi and the original resolution RS1. When the resolution of interest RSi is higher than the original resolution RS1 as in the embodiments of FIGS. 10, 12, and 14, image data of a resolution lower than the resolution of interest RSi may be output. For example, image data having a resolution lower than the original resolution RS1 may be output, or image data having a resolution higher than the original resolution RS1 and lower than the target resolution RSi may be output. Here, when the maximum value U of the number of determinations is 2 or more, the resolution of attention RSi used last may be compared with the resolution of the alternative data. In any case, the alternative data output when the reliability RLb does not satisfy the reliability condition may be binary data or multivalued data, and an image is represented by one color component. It may be data or data representing an image with a plurality of color components. Then, the finally output output image data may be data in various formats including alternative data (for example, PDF data).

(2) The initial value of the resolution of interest RSi in the case where the history information 234 includes similar history information associated with similar feature information having a predetermined similarity relationship with the feature information Ca, instead of the same resolution as the learned resolution RSp. , Various other values determined by using the learned resolution RSp can be adopted. For example, when the noted resolution RSi is updated according to the change amounts dL and dH, the resolution one step before the learned resolution RSp (for example, RSp+dL or RSp-dH) may be adopted. For example, in S230 of FIG. 4, FIG. 6, FIG. 8, FIG. 10, FIG. 12, and FIG. 14, the initial value of the resolution of interest RSi may be set to the resolution one step before the learned resolution RSp. Further, when the initial value determined using the learned resolution RSp is the same as the original resolution RS1, the initial value may be changed to the resolution after one step.

(3) As the configuration of the output image data output by the data output process, various other configurations can be adopted instead of the configuration of the high compression PDF data described above. For example, the high compression PDF data may include the binary image data whose resolution has been converted to the original resolution RS1 without including the drawing command. Further, the size of the character image represented by the high compression PDF data may be different from the size of the character image represented by the scan data. For example, even if the resolution of the binary image data included in the high compression PDF data is different from the resolution of the scan data, the drawing command for enlarging or reducing may be omitted. Further, the gradation number of the character image represented by the high compression PDF data may be different from 2, and may be the same as the gradation number of the scan data, for example. In this case as well, the data size can be reduced by reducing the total number of color components, for example. For example, the high compression PDF data may include monochrome M-gradation image data. Further, the character data added to the high compression PDF data is specified by the character recognition processing of the image data corresponding to the attention resolution different from the highest attention resolution among the plurality of attention resolutions used to calculate the reliability RLb. May be done. For example, the character data Ta specified by the character recognition process of the binary data ia obtained by binarizing the scan data without converting the resolution may be adopted. However, the character data may be omitted. Further, as the data format of the output image data, any other format can be adopted instead of the PDF format (for example, JPG format, PNG format, XPS format). In any case, the gradation number of the character image represented by the output image data may be another gradation number instead of two gradations. Further, the number of gradations of the character image represented by the original image data (for example, scan data) may be another gradation number instead of 256 gradations. In order to reduce the data size of the output image data, the number of gradations of the output image data is preferably smaller than the number of gradations of the original image data.

(4) The storage device that stores the history information 234 may be any other storage device instead of the storage device built in the data processing device 200. For example, an external storage device connected to the communication interface 270 may be adopted.

(5) In the data processing device 200, output processing for reducing the resolution (for example, at least one of FIGS. 4, 6, and 8) and output processing for increasing the resolution (for example, FIGS. 10, 12, and 14). It is preferable that at least one of (1) and (3) is mounted. Further, when a plurality of output processes are implemented, the processor 210 preferably executes the output process selected by the user.

(6) The output device that outputs image data may be a device (for example, a digital camera, a scanner, a smartphone) of a type different from the personal computer such as the data processing device 200 of FIG. 1. Further, the output device may be a part of the printing device or the display device. In addition, a plurality of devices (for example, computers) that can communicate with each other via a network may share the function of the generation process by the generation device, and provide the function of the generation process as a whole (these). The system including the device corresponds to the generation device).

In each of the above embodiments, part of the configuration realized by hardware may be replaced with software, and conversely, part or all of the configuration realized by software may be replaced with hardware. Good. For example, the function of the generation process of S110 of FIG. 3 may be realized by a dedicated hardware circuit.

When some or all of the functions of the present invention are realized by a computer program, the program is provided in a form stored in a computer-readable recording medium (for example, a non-transitory recording medium). be able to. The program can be used while being stored in the same recording medium (computer-readable recording medium) as that provided or provided. The "computer-readable recording medium" is not limited to a portable recording medium such as a memory card or a CD-ROM, but is connected to an internal storage device in the computer such as various ROMs or a computer such as a hard disk drive. External storage may also be included.

Although the present invention has been described above based on the examples and modifications, the above-described embodiments of the present invention are intended to facilitate understanding of the present invention and do not limit the present invention. The present invention can be modified and improved without departing from the spirit and scope of the claims, and the present invention includes equivalents thereof.

200... Data processing device, 210... Processor, 220... Volatile storage device, 230... Nonvolatile storage device, 232... Program, 234... History information, 240... Display Section, 250...operation section, 270...communication interface, 300...multifunction machine, 310...control section, 370...communication interface, 380...scanner section, 390...print execution Department

Claims (27)

An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution larger than the resolution of interest to the reception unit,
Equipped with
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Output device. The output device according to claim 1, wherein
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
The specific resolution is the same as the resolution of the original image data,
Output device. The output device according to claim 1 , wherein
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a lower resolution, and generating the target image data with the updated target resolution,
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data using the attention image data corresponding to the lowest attention resolution among the attention resolutions satisfying the conditions.
Output device. The output device according to claim 3 , wherein
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
When the target resolution is updated at least once, said specific resolution, the I resolution greater der than a resolution less and the target resolution of the original image data, already updated at the end than the pre-update update Is the newest attention resolution among
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution lower than the target resolution to the reception unit,
Equipped with
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Output device. The output device according to claim 5 , wherein
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
The specific resolution is the same as the resolution of the original image data ,
Output device. The output device according to claim 5 , wherein
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a higher resolution, and generating the target image data with the updated target resolution;
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data using the attention image data corresponding to the highest attention resolution among the attention resolutions satisfying the conditions.
Output device. The output device according to claim 7 , wherein
The reference reliability is a reliability represented by using a specific reliability and an allowable change amount,
The specific reliability is specific image data generated by using the original image data, and is a reliability of character recognition of the character image obtained by using the specific image data of a specific resolution representing the character image. Yes,
When the target resolution is updated at least once, said specific resolution, the I smaller resolution der than resolution or more and the target resolution of the original image data, already updated at the end than the pre-update update Is the newest attention resolution among
Output device. The output device according to any one of claims 1 to 8,
The acquisition unit is a character image of interest that is the character image obtained by using the image data of interest, the character recognition of the character image of interest having the same size as the size of the character image when the output image data is expanded. Is obtained as the attention reliability,
Output device. The output device according to any one of claims 1 to 9,
The size of the character image at the time of expanding the output image data is the same as the size of the character image at the time of expanding the original image data,
Output device. It is an output device in any one of Claim 1 to 10, Comprising:
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Output device. The output device according to any one of claims 1 to 11,
The acquisition unit is a target character image that is the character image obtained by using the target image data, and has the same number of gradations as the number of gradations of the character image represented by the output image data. The reliability of character recognition of the target character image is acquired as the target reliability.
Output device. The output device according to any one of claims 1 to 12,
The gradation number of the character image represented by the output image data is smaller than the gradation number of the character image represented by the original image data,
Output device. The output device according to any one of claims 2 , 4, 6, and 8 ,
The output control unit changes the allowable change amount according to the resolution of interest.
Output device. The output device according to any one of claims 3 , 4, 7, and 8 ,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing a feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the initial value of the resolution of interest is determined using the resolution associated with the similar history information,
Output device. The output device according to claim 15, wherein
The original image data represents a character image of a plurality of pages,
The history management unit associates the resolution of the image data of interest used to output the output image data representing the character image of the first page with the feature information representing the feature of the character of the character image of the first page with a specific history. Storing information in the storage device,
When the characteristic information associated with the specific history information has the predetermined similar relationship with the characteristic of the character represented by the character image of the second page, the generation unit may generate the character image of the second page. The initial value of the resolution of interest of the image data of interest is determined using the resolution associated with the specific history information,
Output device. The output device according to claim 15, wherein
The original image data represents a character image of a plurality of pages,
The history management unit associates the resolution of the image data of interest used to output the output image data representing the character image of the first page with the feature information representing the feature of the character of the character image of the first page with a specific history. Storing information in the storage device,
The generation unit determines an initial value of a noticed resolution of noticed image data representing a character image of the second page by using a resolution associated with the specific history information.
Output device. The output device according to any one of claims 1 to 17,
The generation unit generates a plurality of target image data corresponding to a plurality of different target resolutions,
The output control unit, the output image data including character data representing a result of character recognition of the character image obtained by using the attention image data corresponding to the highest attention resolution of the plurality of attention image data, Output,
Output device. The output device according to any one of claims 1 to 18,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing a feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the target resolution has an operation mode for determining using the resolution associated with the similar history information,
The output control unit has an operation mode of outputting the output image data by using the attention image data of the attention resolution regardless of whether or not the condition is satisfied.
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution larger than the resolution of interest to the reception unit,
Equipped with
The acquisition unit is a character image of interest that is the character image obtained by using the image data of interest, the character recognition of the character image of interest having the same size as the size of the character image when the output image data is expanded. Is obtained as the attention reliability,
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution lower than the target resolution to the reception unit,
Equipped with
The acquisition unit is a character image of interest that is the character image obtained by using the image data of interest, the character recognition of the character image of interest having the same size as the size of the character image when the output image data is expanded. Is obtained as the attention reliability,
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution larger than the resolution of interest to the reception unit,
Equipped with
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a lower resolution, and generating the target image data with the updated target resolution,
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data by using the attention image data corresponding to the lowest attention resolution among the attention resolutions satisfying the conditions,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing a feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the initial value of the resolution of interest is determined using the resolution associated with the similar history information,
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution lower than the target resolution to the reception unit,
Equipped with
The generation unit and the acquisition unit,
Updating the target resolution by the generation unit to a higher resolution, and generating the target image data with the updated target resolution;
Acquiring the attention reliability of the attention image data corresponding to the updated attention resolution by the acquisition unit;
Is repeated until the attention reliability corresponding to the updated attention resolution does not satisfy the condition,
The output control unit outputs the output image data by using attention image data corresponding to the highest attention resolution among the attention resolutions satisfying the conditions,
A history management unit that stores history information in which a resolution of interest satisfying the condition and feature information representing a feature of the character represented by the character image are stored in a storage device;
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the initial value of the resolution of interest is determined using the resolution associated with the similar history information,
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution larger than the resolution of interest to the reception unit,
A history management unit that stores history information that associates a resolution of interest satisfying the condition and feature information that represents the feature of the character represented by the character image in a storage device;
Equipped with
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the target resolution has an operation mode for determining using the resolution associated with the similar history information,
The output control unit has an operation mode of outputting the output image data by using the attention image data of the attention resolution regardless of whether or not the condition is satisfied.
Output device. An output device for outputting image data representing a character image including characters,
A generation unit that generates, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition unit that acquires the reliability of attention, which is the reliability of character recognition of the character image obtained using the target image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control unit that outputs image data having a resolution lower than the target resolution to the reception unit,
A history management unit that stores history information that associates a resolution of interest satisfying the condition and feature information that represents the feature of the character represented by the character image in a storage device;
Equipped with
The generation unit stores similarity history information, which is history information associated with feature information having a predetermined similarity relationship with a feature of a character represented by a character image of the original image data to be processed, in the storage device. In the case of, the target resolution has an operation mode for determining using the resolution associated with the similar history information,
The output control unit has an operation mode of outputting the output image data by using the attention image data of the attention resolution regardless of whether or not the condition is satisfied.
Output device. A computer program for causing a computer to output image data representing a character image including characters,
A generation function of generating, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest lower than the resolution of the original image data;
An acquisition function for acquiring attention reliability, which is reliability of character recognition of the character image obtained using the attention image data,
When a condition including that the attention reliability is equal to or higher than a reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control function of outputting image data having a resolution higher than the target resolution to the reception unit,
To the computer ,
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Computer program. A computer program for causing a computer to output image data representing a character image including characters,
A generation function for generating, using the original image data representing the character image, the image data of interest representing the character image at a resolution of interest higher than the resolution of the original image data;
An acquisition function for acquiring attention reliability, which is reliability of character recognition of the character image obtained using the attention image data,
When a condition including that the attention reliability is less than the reference reliability is satisfied, output image data is output to a reception unit that receives data by using the attention image data having the attention resolution, and the condition is satisfied. If not, an output control function of outputting image data having a resolution lower than the target resolution to the reception unit,
To the computer ,
The output image data includes the target image data of the target resolution and a command for enlarging or reducing the character image represented by the target image data.
Computer program.

Images