Image processing device, image formation device, program, and image correction method

Abstract

To provide an image processing device, an image formation device, a program, and an image correction method that perform appropriate image correction on input images in accordance with various OCR software, thereby improving the accuracy of OCR processing.SOLUTION: An image processing device includes an image capturing unit 40 that captures an image and a processing unit 30 that includes an OCR pre-processing unit 30-1. The OCR pre-processing unit includes: an OCR software information setting unit 31 configured to set OCR software information according to OCR software to be used later; an image correction unit 32 configured to perform image correction on the image captured by the image capturing unit 40 in accordance with the OCR software information set in the OCR software information setting unit 31; and an image data output unit 33 configured to output the image corrected by the image correction unit 32.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to an image processing apparatus, an image forming apparatus, a program, and an image correction method. [Background technology]

[0002] Conventionally, OCR (Optical Character Recognition) processing technology is known that automatically extracts character information from a scanned image.

[0003] Patent Document 1 discloses a technique for automatically setting processing conditions for OCR software installed on a user terminal, with the aim of improving user convenience. Summary of the Invention [Problem to be solved by the invention]

[0004] However, with conventional technology, it was only possible to set processing condition information for the output results of specific OCR software that had been pre-installed on a user terminal, etc., and there was a problem in that it was not possible to improve the accuracy of OCR processing for various general-purpose OCR software.

[0005] The present invention has been made in view of the above, and has as its object to perform appropriate image correction on an input image in accordance with various OCR software, thereby improving the accuracy of OCR processing. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the object, the present invention is characterized by comprising an image capturing unit that captures an image, an OCR software information setting unit that sets OCR software information according to subsequent OCR software, an image correction unit that performs image correction on the image captured by the image capturing unit according to the OCR software information set in the OCR software information setting unit, and an image data output unit that outputs the image corrected by the image correction unit. [Effects of the Invention]

[0007] According to the present invention, it is possible to perform appropriate image correction on an input image in accordance with various OCR software, thereby improving the accuracy of OCR processing. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of a configuration of a reading device according to the first embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of the configuration of a control block of the reading device. [Figure 3] FIG. 3 is a diagram showing a schematic configuration for performing OCR preprocessing. [Figure 4] FIG. 4 is a flowchart showing the flow of OCR preprocessing. [Figure 5] FIG. 5 is a diagram showing an example of a software selection screen on the operation panel. [Figure 6] FIG. 6 is a diagram showing an example of a function selection screen on the operation panel. [Figure 7] FIG. 7 is a diagram showing an example in which OCR software information, a color removal / seal imprint removal function, and image correction are linked together. [Figure 8] FIG. 8 is a diagram illustrating an image relating to an example of a failed OCR process. [Figure 9] FIG. 9 is a diagram illustrating an example in which the OCR processing result changes depending on the OCR processing method of the OCR software. [Figure 10] FIG. 10 is a diagram illustrating an example in which the OCR processing result changes when chromatic colors are lightened by image correction. [Figure 11] FIG. 11 is a diagram illustrating the difference in character quality and file size due to compression. [Figure 12] FIG. 12 is a diagram illustrating a low-compression image correction method. [Figure 13]FIG. 13 is a diagram showing an example of a normal compression table and a low compression compression table. [Figure 14] FIG. 14 is a diagram schematically illustrating a configuration for performing OCR preprocessing according to the second embodiment. [Figure 15] FIG. 15 is a diagram illustrating the difference in character quality and file size depending on the resolution. [Figure 16] FIG. 16 is a diagram schematically illustrating a configuration for performing OCR preprocessing according to the third embodiment. [Figure 17] FIG. 17 is a diagram illustrating image correction using an invisible image. [Figure 18] FIG. 18 is a diagram showing the light absorption characteristics of color materials. [Figure 19] FIG. 19 is a diagram schematically illustrating a configuration for performing OCR preprocessing according to the fourth embodiment. [Figure 20] FIG. 20 is a diagram schematically illustrating a configuration for performing OCR preprocessing according to the fifth embodiment. [Figure 21] FIG. 21 is a diagram illustrating automatic linking of image correction by AI according to the sixth embodiment. [Figure 22] FIG. 22 illustrates an example of a configuration of an image forming apparatus according to the seventh embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of an image processing apparatus, an image forming apparatus, a program, and an image correction method will be described in detail with reference to the accompanying drawings.

[0010] (First embodiment) FIG. 1 is a diagram showing an example of the configuration of a reading device 1 according to a first embodiment. Hereinafter, the subject may be referred to as a reading target. The reading target is an area in which black characters or the like are printed and an object to be removed, such as a seal impression, overlaps the black character area. As an example, a paper document such as various certificates, documents, or slips bearing a seal is shown, but the object to be removed is not limited to a seal impression. Furthermore, the reading target is not limited to a paper document.

[0011] As shown in FIG. 1, a main body 10 of a reading device 1, which is an image processing device, has a contact glass 11 on its top surface, and an image capturing unit 40 (see FIG. 2) inside the main body 10. Inside the main body 10, a light source 13, a first carriage 14, a second carriage 15, a lens unit 16, an image sensor 17, etc. are provided. The first carriage 14 has the light source 13 and a reflecting mirror 14-1, and the second carriage 15 has reflecting mirrors 15-1 and 15-2. The main body 10 also has a control board. The control board is a control unit 300 shown in FIG. 2, which controls the entire reading device 1.

[0012] The control board moves the first carriage 14 and the second carriage 15, irradiating them with light from the light source 13, and sequentially reading the light reflected from the object to be read placed on the contact glass 11 with the image sensor 17. The light reflected from the object to be read by the light from the light source 13 is reflected by the mirror 14-1 of the first carriage 14 and the mirrors 15-1 and 15-2 of the second carriage 15, and enters the lens unit 16, and the light emitted from the lens unit 16 forms an image on the image sensor 17. The image sensor 17 receives the light reflected from the object to be read and outputs an image signal. The image sensor 17 is, for example, a CCD (Charge Coupled Device) or CMOS (Complementary MOS) image sensor, and corresponds to a reading unit that reads the image of the object to be read.

[0013] The reference white plate 12 is a member for white correction.

[0014] The reading device 1 shown in Fig. 1 is equipped with an ADF (Automatic Document Feeder) 20. The ADF 20 opens when one side is lifted up, exposing the surface of the contact glass 11. The user places the object to be read on the contact glass 11, lowers the ADF 20, and presses the ADF 20 against the surface of the contact glass 11 from behind the object to be read. Then, when a scan start button is pressed, the first carriage 14 and the second carriage 15 are driven in the main scanning direction and the sub-scanning direction, and the entire object to be read is read.

[0015] In addition to the method of placing the object to be read on the contact glass 11 and reading it, the object to be read can also be read in the following way. The ADF 20 is also capable of reading the object to be read using a sheet-through method. In the ADF 20, a pickup roller 22 separates a stack of objects to be read from a tray 21 of the ADF 20 one by one, and by controlling various conveyance rollers 24, etc., one or both sides of the object to be read conveyed along a conveyance path 23 are read and the object is discharged to an output tray 25.

[0016] Reading of the object to be read by the ADF 20 using the sheet-through method is performed through the reading window 19. In this example, the first carriage 14 and the second carriage 15 are moved to and fixed at a predetermined home position, and when the object to be read passes between the reading window 19 and the background portion 26, the image is read by irradiating the surface of the object to be read that faces the reading window 19 with light from the light source 13. The reading window 19 is a slit-shaped reading window provided in part of the contact glass 11. The background portion 26 is a background member.

[0017] When ADF 20 is used to read both sides of a document, after the document passes through reading window 19, the back side is read by reading module 27, a second reading means provided on the back side of the document. Reading module 27 has an irradiation unit including a light source and a contact image sensor, which is a second reading unit, and reads the reflected light of the light irradiated onto the second side with the contact image sensor. This light source may also be provided with a visible light source and a near-infrared light source, so that both visible and near-infrared images can be read. Background member 28 is a density reference member.

[0018] Next, the configuration of the control block of the reading device 1 will be described.

[0019] FIG. 2 is a diagram showing an example of the configuration of the control block of the reading device 1. As shown in FIG. 2, the reading device 1 has a control unit 300, an operation panel 301, various sensors 302, a scanner motor 303, various motors 304 for the transport path, a drive motor 305, an output unit 306, and an image capturing unit 40. In addition, various control objects are connected. The various sensors 302 are sensors that detect the object to be read. The scanner motor 303 is a motor that drives the first carriage 14 and the second carriage 15 of the main body 10. The various motors 304 for the transport path are various motors provided in the ADF 20. The output unit 306 corresponds to an output interface for outputting image data to an external device. The output interface may be an interface such as a USB or a communication interface that connects to a network.

[0020] The operation panel 301 is, for example, a touch panel type liquid crystal display device. The operation panel 301 accepts input operations such as various settings and reading execution (scan start) from the user via operation buttons, touch input, etc., and transmits corresponding operation signals to the control unit 300. The operation panel 301 also displays various display information from the control unit 300 on the display screen.

[0021] For example, the operation panel 301 displays a software selection screen with selection buttons for selecting the name of OCR software, and when an input operation is performed on the selection button, it instructs the control unit 300 to make the selection. The operation panel 301 also displays a function setting screen with selection buttons for selecting the functions of the OCR software selected on the software selection screen, and when an input operation is performed on the selection button, it instructs the control unit 300 to make the selection.

[0022] The image capturing unit 40 has a light source unit 401, a sensor chip 402, an amplifier 403, an A / D 404, an image correction processing unit 405, a frame memory 406, an output control circuit 407, and an I / F circuit 408, and an image read from a reading target is output frame by frame from the output control circuit 407 to the control unit 300 via the I / F circuit 408. Each sensor chip 402 is a pixel sensor of the image sensor 17. The light source unit 401 is the light source 13.

[0023] The image capturing unit 40 is driven by the controller 307. For example, the image capturing unit 40 turns on the light source unit 401 based on a lighting signal from the controller 307, and irradiates the object to be read with light at a set timing. The image capturing unit 40 also converts the light from the object to be read, which is imaged on the sensor surface of the image sensor 17, into an electrical signal in each sensor chip 402 and outputs the signal.

[0024] The image capturing unit 40 amplifies pixel signals output from each sensor chip 402 using an amplifier 403, converts the analog signals to digital signals using an A / D 404, and outputs pixel level signals. The image correction processing unit 405 performs image correction processing on the output signals from each pixel. For example, the image correction processing unit 405 performs shading correction on the output signals from each pixel.

[0025] After the image correction process, each data is stored in a frame memory 406 , and the stored read image is transferred to the control unit 300 via an output control circuit 407 and an I / F circuit 408 .

[0026] The control unit 300 includes a CPU (Central Processing Unit), memory, etc., and the CPU controls the entire device to perform a reading operation on the reading target and an OCR pre-processing on the read image obtained by the reading operation, prior to the OCR processing.

[0027] The control unit 300 has a processing unit 30. The processing unit 30 can be realized by a CPU executing a predetermined program, or alternatively, by hardware such as an ASIC (Application Specific Integrated Circuit).

[0028] Fig. 3 is a diagram schematically showing a configuration for executing OCR preprocessing, and Fig. 4 is a flowchart showing the flow of OCR preprocessing. Image capturing unit 40 includes light source 13 and image sensor 17 described in Fig. 1 and Fig. 2. Image capturing unit 40 irradiates the object to be read with light from light source 13, receives reflected light from the object to be read with image sensor 17, and outputs an image (an RGB image, for example).

[0029] 3, the processing unit 30 includes an OCR preprocessing unit 30-1, which includes an OCR software information setting unit 31, an image correction unit 32, and an image data output unit 33.

[0030] The OCR software information setting unit 31 sends OCR software information corresponding to the functions of the subsequent OCR software to the image correction unit 32 (step S1 in FIG. 4). Note that the subsequent OCR software is not limited to the OCR software installed in the reading device 1, which is an image processing device. For example, it refers to all OCR software that uses images generated by the reading device 1, which is an image processing device, such as OCR software on the cloud or OCR software installed on a PC.

[0031] For example, the OCR software information setting unit 31 displays a software selection screen on the operation panel 301.

[0032] FIG. 5 is a diagram showing an example of a software selection screen D1 on the operation panel 301. As shown in FIG. 5, the OCR software information setting unit 31 displays the software selection screen D1 on the operation panel 301. The display example shown in FIG. 5 shows an example in which the name of OCR software "Receipt Management XX" is selected and the OK button is pressed. By accepting the selection of the OCR software name on the software selection screen D1, the OCR software information setting unit 31 sets OCR software information according to the selected OCR software name. In this way, the user does not need to understand the processing of the OCR software, and can perform appropriate image correction simply by selecting the OCR software name.

[0033] Also, for example, the OCR software information setting unit 31 displays a function selection screen on the operation panel 301.

[0034] FIG. 6 is a diagram showing an example of a function selection screen D2 on the operation panel 301. As shown in FIG. 6, the OCR software information setting unit 31 displays the function selection screen D2 on the operation panel 301. The display example shown in FIG. 6 shows an example in which OCR software functions "without color removal / shade removal" and "without AI" are selected and the OK button is pressed. By accepting the selection of the OCR software functions on the function selection screen D2, the OCR software information setting unit 31 sets OCR software information according to the selected OCR software functions. In this way, the user can perform appropriate image correction simply by selecting a function without having to understand the OCR software's processing.

[0035] The image correction unit 32 performs OCR preprocessing for increasing the recognition rate of OCR processing or normal image processing for improving image quality, which is set according to the OCR software information sent from the OCR software information setting unit 31, on the image acquired by the image capturing unit 40, and sends the image to the image data output unit 33 (step S2 in FIG. 4). Examples of OCR preprocessing include the presence or absence of AI (artificial intelligence), color removal, seal imprint removal, and image correction.

[0036] Fig. 7 is a diagram showing an example in which OCR software information, a color removal / imprint removal function, and image correction are linked together. As shown in Fig. 7, the image correction unit 32 stores the OCR software information, the presence or absence of AI (Artificial Intelligence), the color removal / imprint removal function, and the content of image correction that defines what kind of image correction should be performed, in association with each other.

[0037] In addition, OCR processing using AI involves using AI to make inferences based on data that has been learned up to now (teaching data) and then carrying out OCR processing. By using AI to make inferences based on data that has been learned up to now (teaching data) and then carrying out OCR processing in this way, the OCR software using AI can be corrected to have image quality close to that of the teaching data, thereby maintaining the accuracy of the AI ​​OCR.

[0038] 7, the image correction unit 32 has a function to turn off pre-processing for OCR. This is because when inference is performed using AI based on data that has been learned up to now (teaching data) and OCR processing is performed, it is possible that unconventional image processing will be performed, resulting in image quality that differs from that of the teaching data, and there is a risk that the OCR processing will fail during inference.

[0039] The image correction unit 32 uses the OCR software information sent from the OCR software information setting unit 31 and refers to the data stored as described above, thereby switching the image correction process.

[0040] The image data output unit 33 outputs the image sent from the image correction unit 32, such as by transmitting the data or displaying it on the operation panel 301 (step S3 in FIG. 4).

[0041] Below, we will explain the OCR pre-processing that is executed by the image correction unit 32, which uses the OCR software information sent from the OCR software information setting unit 31 to increase the recognition rate of the OCR processing on the image acquired by the image capturing unit 40.

[0042] OCR processing technology, which automatically extracts text information from scanned images, is well known. There are many types of OCR software that can achieve this. Each type of OCR software also employs different processing methods. OCR processing methods include software with functions for removing print impressions and colors, software without these functions, and software with recognition functions that utilize AI (Artificial Intelligence).

[0043] Depending on the OCR processing method, the image quality that facilitates successful OCR processing varies. For example, software equipped with imprint removal or color removal functions requires image quality that accurately identifies the text and facilitates the color removal / imprint removal functions when there is color content surrounding the text. Software without imprint removal or color removal functions also requires image quality that eliminates the color content (corrected to white during binarization) when there is color content surrounding the text, or image quality that allows the text to be identified during image / text detection. The process of generating image quality that facilitates successful OCR processing, such as this, is called OCR preprocessing. On the other hand, software that utilizes AI (artificial intelligence), requires image quality that has undergone standard image processing to closely match the training data.

[0044] As described above, the image quality required for easy character recognition varies depending on the OCR processing method used by the OCR software. Therefore, in this embodiment, the OCR preprocessing of the image captured by the image capturing unit 40 is changed depending on the OCR software used in the subsequent stage, thereby improving the accuracy of the OCR processing.

[0045] Here, an example of a failed OCR process will be described.

[0046] Fig. 8 is a diagram explaining an example of an image where OCR processing has failed, and Fig. 9 is a diagram explaining an example where the OCR processing results change depending on the OCR processing method used by the OCR software. A typical example of an image where OCR processing has failed is an image that contains a mixture of black text and colored content (seals, ruled lines, background patterns, etc.), as shown in Fig. 8.

[0047] Commonly performed pre-processing steps for OCR to improve the recognition rate include color removal / imprint removal, picture / character area determination, binarization, and character recognition.

[0048] As shown in Figure 9(a), if software A is equipped with color removal / seal impression removal functions, in the case of the image shown in Figure 8, the character area can be accurately determined in the picture / character area determination process, and the color content (seal, ruled lines, background pattern, etc.) can be eliminated in the binarization process, leaving only the black character area, so the OCR process is successful and character recognition becomes possible.

[0049] On the other hand, as shown in Figure 9(b), software B, which does not have the color removal / seal impression removal function, can determine the character area in the picture / character area determination process for the image shown in Figure 8, but the color content (seal, ruled lines, background patterns, etc.) does not disappear in the binarization process, and the color content (seal, ruled lines, background patterns, etc.) around the black characters interferes with character recognition, so the OCR process fails and character recognition becomes impossible.

[0050] Furthermore, as shown in Figure 9(c), in the case of software C that does not have the color removal / seal impression removal function, in the case of the image shown in Figure 8, the image is judged to be a picture during the picture / character area determination process because the density of the color content (seal, ruled lines, background patterns, etc.) is too high, and OCR processing is not performed, so the OCR processing fails and character recognition is not possible.

[0051] As mentioned above, if the color content (seals, lines, patterns, etc.) around black text is highly concentrated, the text may be mistaken for a picture. Furthermore, the color content (seals, lines, patterns, etc.) may not disappear during binarization. While removing the color in an image correction process prior to OCR processing is one solution to this problem, it also creates a new problem: the resulting image may appear significantly different from the original.

[0052] Therefore, in this embodiment, the image correction unit 32 changes the image processing of the image acquired by the image capturing unit 40 in accordance with the OCR software (by reducing chromatic colors while leaving color information intact), thereby preventing erroneous recognition during the picture / text area determination process and binarization process without significantly impairing the appearance of the original image and improving the accuracy of the OCR process. This will be described in detail below.

[0053] FIG. 10 is a diagram illustrating an example in which the OCR processing result changes when chromatic colors are lightened by image correction.

[0054] As shown in Figure 10(a), software A equipped with color removal / seal impression removal functions is used in the case of the image shown in Figure 8. If the image quality is suitable for OCR processing, the character area can be accurately determined in the picture / character area determination process, and the color content (seal, ruled lines, background patterns, etc.) can be eliminated in the binarization process, leaving only the black character area, so that the OCR process is successful and characters can be recognized.

[0055] On the other hand, as shown in Figure 10(b), software B, which does not have the color removal / seal impression removal function, can determine the character area in the picture / character area determination process when the chromatic colors are lightly corrected to an image quality suitable for OCR processing in the case of the image shown in Figure 8, and the color content (seal, ruled lines, background patterns, etc.) can be eliminated in the binarization process, leaving only the black character area, so the OCR process is successful and characters can be recognized.

[0056] Furthermore, as shown in Figure 10(c), software C, which does not have a color removal / seal impression removal function, can determine the character area in the picture / character area determination process when the image shown in Figure 8 is used and the chromatic colors are lightly corrected to an image quality suitable for OCR processing. In the binarization process, the color content (seals, lines, patterns, etc.) disappears, leaving only the black character area, so the OCR process is successful and characters can be recognized.

[0057] As described above, the image correction unit 32 changes the image processing of the image acquired by the image capturing unit 40 in accordance with the OCR software (for example, by lightening chromatic colors while leaving color information intact), thereby preventing erroneous recognition during the picture / text area determination process and binarization process without significantly impairing the appearance of the original image and improving the accuracy of the OCR process.

[0058] In other words, the image correction unit 32 changes the image processing of the image acquired by the image capturing unit 40 depending on the OCR software (by lightening chromatic colors while leaving color information intact), thereby preventing OCR software without color removal or seal impression removal from being unable to recognize characters and improving the accuracy of OCR processing.

[0059] In the case of accounting forms, the chromatic color around the characters is often the red of a seal. Therefore, the image correction unit 32 may correct only red hues. In this case, colors other than red can be output in the same color as before.

[0060] In other words, the image correction unit 32 changes the image processing of the image acquired by the image capturing unit 40 (by lightening the red color) depending on the OCR software, thereby preventing OCR software without color removal or seal impression removal from being unable to recognize characters, while outputting colors other than red at the same level as before, in accounting forms with many seals, etc., thereby improving the accuracy of OCR processing.

[0061] As described above, according to this embodiment, by changing the image quality processing depending on the OCR software in the subsequent stage, it is possible to provide the OCR software with the most suitable image quality and improve the accuracy of the OCR processing.

[0062] In this embodiment, the image correction unit 32 changes the image processing of the image acquired by the image capturing unit 40 (diluting chromatic colors while leaving color information intact) depending on the OCR software, but this is not limited to this.

[0063] For example, the image correction unit 32 may change the image processing of the image acquired by the image capturing unit 40 (switch the compression level) depending on the OCR software.

[0064] Figure 11 explains the difference in character quality and file size due to compression, and Figure 12 explains a low-compression image correction method. For example, some OCR software has strict file size limits. OCR software with strict file size limits can eliminate the risk of size errors by using high-compression image correction. On the other hand, there is also OCR software with a looser file size limit.

[0065] As shown in Figure 11(a), a highly compressed image has poor text quality but a small file size, while as shown in Figure 11(b), a low-compression image has good text quality but a large file size.

[0066] Therefore, for example, in the case of OCR software with a loose upper limit on file size, the image correction unit 32 performs low-compression image correction to provide an image with high character quality and improve the accuracy of OCR processing.

[0067] As shown in Figure 12, in the case of OCR software with a loose upper limit on file size, the image correction unit 32 can generate an image with the best character quality and improve the accuracy of the OCR processing by turning off the subsampling process that thins out the data and performing compression using a low-compression compression table.

[0068] 12, in the case of OCR software with a loose upper limit on file size, the image correction unit 32 can generate images with high character quality and improve the accuracy of OCR processing by turning off subsampling processing, which thins out data, and performing compression using a normal compression table. In this case, it is sufficient to change the subsampling setting, and the effort required for the change is small.

[0069] 12, in the case of OCR software with a loose upper limit on file size, the image correction unit 32 generates an image with high character quality and improves the accuracy of OCR processing by turning on subsampling processing, which thins out data, and performing compression using a low-compression compression table. In this case, it is possible to generate an image with high character quality and improve the accuracy of OCR processing while being compatible with OCR software or image viewers that do not support subsampling off.

[0070] Here, low compression refers to a situation in which more than half of the elements in the compression table have smaller values ​​than in normal compression. Here, FIG. 13 is a diagram showing an example of a normal compression table and a low-compression compression table. For example, an example of a normal compression table is shown in FIG. 13(a), and an example of a low-compression compression table is shown in FIG. 13(b). In the low-compression compression table in FIG. 13(b), the values ​​that are smaller than those in the normal compression table in FIG. 13(a) are highlighted. As shown in FIG. 13(b), it can be seen that in the low-compression compression table, more than half of the values ​​(53 values, i.e., more than half, i.e., 32 or more values) are smaller than those in the normal compression table. However, the combination of the normal compression table and the low-compression compression table shown in FIG. 13 is merely one example.

[0071] As described above, image correction unit 32 can generate an image that matches the upper limit of the file size set by the OCR software by changing the image processing of the image acquired by image capture unit 40 (switching the compression level) depending on the OCR software. Also, for example, image correction unit 32 may change the image processing of the image acquired by image capture unit 40 (switching the compression level) depending on the OCR software.

[0072] In response to an instruction from the downstream OCR software, the OCR software information setting unit 31 may send OCR software information corresponding to the OCR software to the image correction unit 32. By switching the image correction process in response to an instruction from the downstream OCR software in this way, it is possible to reduce the effort required for user settings.

[0073] (Second embodiment) Next, a second embodiment will be described.

[0074] The second embodiment differs from the first embodiment in that the image correction unit 32 changes the image processing (switches the resolution) of the image acquired by the image capturing unit 40 depending on the OCR software. In the following description of the second embodiment, the same parts as in the first embodiment will be omitted, and only the parts that differ from the second embodiment will be described.

[0075] Here, FIG. 14 is a diagram showing a schematic configuration for executing OCR preprocessing according to the second embodiment.

[0076] As shown in FIG. 14, the OCR software information setting unit 31 sends OCR software information corresponding to the function of the subsequent OCR software to at least one of the image capturing unit 40 and the image correcting unit 32.

[0077] The image correction unit 32 uses the OCR software information sent from the OCR software information setting unit 31 to perform OCR preprocessing on the image acquired by the image capturing unit 40 to increase the recognition rate of the OCR process, or normal image processing to improve image quality, and sends the image to the image data output unit 33.

[0078] The image capturing unit 40 changes the resolution using the OCR software information sent from the OCR software information setting unit 31. The image capturing unit 40 can change the resolution by, for example, switching the reading resolution of the image sensor or switching the transport speed of the object to be read.

[0079] Here, Figure 15 is a diagram explaining the difference in character quality and file size depending on the resolution. For example, some OCR software has strict upper limits on file size. For OCR software with strict upper limits on file size, the risk of size errors can be eliminated by capturing low-resolution images and correcting the images. On the other hand, there is also OCR software with a loose upper limit on file size. For OCR software with a loose upper limit on file size, capturing high-resolution images and correcting the images can provide images with high character quality and improve the accuracy of the OCR process.

[0080] In recent years, there has been a growing demand for not only compatibility with OCR software but also compliance with the Electronic Bookkeeping Act, which sometimes requires a scan resolution of 200 dpi or higher. In such cases, this system can capture high-resolution images of 200 dpi or higher while still meeting the requirements of OCR software.

[0081] Thus, according to this embodiment, at least one of the image capturing unit 40 and the image correction unit 32 changes the image processing (switches the resolution) of the image acquired by the image capturing unit 40 depending on the OCR software, thereby making it possible to generate an image with even higher resolution.

[0082] In this embodiment, the OCR software information setting unit 31 may send OCR software information corresponding to the OCR software to at least one of the image capturing unit 40 and the image correction unit 32 in response to instructions from the downstream OCR software.

[0083] (Third embodiment) Next, a third embodiment will be described.

[0084] The third embodiment differs from the first embodiment in that the image capturing unit 40 has a visible image capturing unit 40-1 and an invisible image capturing unit 40-2. In the following description of the third embodiment, the description of the same parts as in the first embodiment will be omitted, and only the parts that differ from the third embodiment will be described.

[0085] 16 is a diagram illustrating a configuration for performing OCR preprocessing according to the third embodiment. As shown in FIG. 16, the image capturing unit 40 includes a visible image capturing unit 40-1 and an invisible image capturing unit 40-2.

[0086] The light source 13 includes a visible light source 13-1 and an invisible light source 13-2. The image sensor 17 includes a first image sensor 17-1 that receives reflected visible light irradiated onto the object to be read and outputs an image, and a second image sensor 17-2 that receives reflected near-infrared light irradiated onto the object to be read and outputs an image. An image received by irradiating the object with visible light is called a visible image, and an image received by irradiating the object with near-infrared light is called a near-infrared image. Although the light source 13 includes separate visible light source 13-1 and invisible light source 13-2, they may be combined into a single light source.

[0087] The first image sensor 17-1 and the second image sensor 17-2 may be configured as a single image sensor, or each may be provided as a separate image sensor, as long as they are capable of outputting a visible image and a near-infrared image, respectively.

[0088] The visible image capturing unit 40-1 irradiates the object to be read with light from the visible light source 13-1, receives the light reflected from the object with a visible image sensor 17-1, and outputs a visible image (e.g., an RGB image). The invisible image capturing unit 40-2 irradiates the same object to be read with light from the invisible light source 13-2, receives the light reflected from the object with a near-infrared image sensor 17-2, and outputs a near-infrared image (NIR image).

[0089] The image capturing unit 40 can simultaneously capture both a visible image and a near-infrared image from the same target. Note that if the target is the same, the image capturing unit 40 does not need to simultaneously capture both a visible image and a near-infrared image, and they may be captured at different times as long as the positions of the respective target images are the same.

[0090] 17 is a diagram illustrating image correction using an invisible image. The example shown in FIG. 17 is an example in which the image correction unit 32 corrects the visible image acquired by the visible image capturing unit 40-1 using the invisible image and the OCR software information sent from the OCR software information setting unit 31.

[0091] Incidentally, the black ink, black toner, and black pencils used for black text contain carbon. Carbon has the characteristic of absorbing light in the visible and infrared regions. Therefore, it can be read as black in both the visible and infrared regions. On the other hand, color ink and CMY toner have the characteristic of transmitting light in the infrared region. This characteristic is shown in Figure 18. The horizontal axis represents the wavelength of light, and the vertical axis represents the light absorption rate. In Figure 18, black refers to the color of materials containing carbon as described above, gray refers to the color printed by thinning out gray ink or black toner, and color mixture refers to the color created by mixing CMY colors (hereinafter referred to as color mixture black). Color mixture black absorbs light in the visible region but transmits light in the near-infrared region, so it can be read as white like a blank sheet of paper in the near-infrared region.

[0092] As shown in Fig. 17(a), the image read in the visible region is an image (visible image) in which black text and color content (seals, ruled lines, background patterns, etc.) are mixed together. More specifically, as shown in Fig. 17(a), the image read in the visible region is an image (visible image) in which color content (seals, ruled lines, background patterns, etc.) are superimposed on black text.

[0093] On the other hand, as shown in FIG. 17(b), the image read in the infrared region is an image of only black characters (invisible image).

[0094] 17(c), the image correction unit 32 determines the position of the black characters in the invisible image and performs correction to return the overlaid black characters to their original black color. This allows the original character color to be restored even if the color of the characters has been changed by a seal or the like, and the character parts will not be erased by the seal impression removal / color removal process of the OCR software, improving the accuracy of the OCR process.

[0095] In this way, according to this embodiment, the results of the invisible image can be used to perform image correction to return black text, etc., that has changed color due to overlaying color content (seals, lines, patterns, etc.), back to black, thereby improving the accuracy of OCR processing.

[0096] In the first embodiment, the image correction unit 32 reduces chromatic colors while leaving color information intact, or switches the compression level, in the second embodiment, switches the resolution, and in the third embodiment, performs correction using an invisible image, depending on the OCR software information of the OCR software. However, the image correction unit 32 is not limited to these individual image corrections, and may also use a combination of each image correction. This makes it possible to generate an image optimal for OCR processing depending on the OCR software.

[0097] (Fourth embodiment) Next, a fourth embodiment will be described.

[0098] The fourth embodiment differs from the third embodiment in that the image correction unit 32 switches the operation of the image capturing unit 40 (visible image capturing unit 40-1, invisible image capturing unit 40-2) depending on the OCR software in the subsequent stage. In the following description of the fourth embodiment, the description of the same parts as in the third embodiment will be omitted, and only the parts that differ from the third embodiment will be described.

[0099] Here, FIG. 19 is a diagram showing a schematic configuration for executing OCR preprocessing according to the fourth embodiment.

[0100] As shown in FIG. 19, the OCR software information setting unit 31 sends OCR software information corresponding to the function of the subsequent OCR software to at least one of the image capturing unit 40 and the image correcting unit 32.

[0101] The image correction unit 32 uses the OCR software information sent from the OCR software information setting unit 31 to perform OCR preprocessing on the image acquired by the image capturing unit 40 to increase the recognition rate of the OCR process, or normal image processing to improve image quality, and sends the image to the image data output unit 33.

[0102] The image acquired by the image capturing unit 40 is an example, and the image corrected by the image correcting unit 32 does not have to be a captured image. For example, image correction may be performed on an image stored in an internal storage or an image sent via a network.

[0103] The image capturing unit 40 may switch its operation depending on the OCR software information sent from the OCR software information setting unit 31. Examples of the mode will be described below.

[0104] (When the visible image capturing unit 40-1 and the invisible image capturing unit 40-2 are operated) For example, when a specific OCR software is used, the image capturing unit 40 operates the visible image capturing unit 40-1 and the invisible image capturing unit 40-2. This allows the accuracy of the OCR processing to be improved by performing correction using invisible images for the OCR software, which improves accuracy by performing image correction using invisible images.

[0105] (When only the visible image capturing unit 40-1 is operated) For example, when specific OCR software is used, only the visible image capturing unit 40-1 is operated in the image capturing unit 40. This makes it possible to remove unnecessary invisible components when capturing a visible image, and therefore, compared to when both the visible image capturing unit 40-1 and the invisible image capturing unit 40-2 are operated, it is possible to prevent the visible image from being mixed with invisible components, and improve the image quality of the visible image.

[0106] (When only the invisible image capturing unit 40-2 is operated) For example, when specific OCR software is used, the image capturing unit 40 operates only the invisible image capturing unit 40-2. This changes the appearance due to capturing invisible images, but improves the accuracy of OCR processing without performing special image processing.

[0107] Thus, according to this embodiment, by using the OCR software information from the OCR software information setting unit 31 to operate the visible image capturing unit 40-1 or the invisible image capturing unit 40-2, or both, an image can be captured that matches the OCR software.

[0108] (Fifth embodiment) Next, a fifth embodiment will be described.

[0109] The fifth embodiment differs from the first embodiment in that the image correction unit 32 includes a common correction unit 32-1 and an OCR software information switching correction unit 32-2. In the following description of the fifth embodiment, the description of the same parts as in the first embodiment will be omitted, and only the differences from the fifth embodiment will be described.

[0110] 20 is a diagram illustrating a configuration for performing OCR preprocessing according to the fifth embodiment. As shown in FIG. 20, the image correction unit 32 includes a common correction unit 32-1 and an OCR software information switching correction unit 32-2.

[0111] The common correction unit 32-1 executes correction processes common to various types of OCR software, such as noise removal and background removal.

[0112] The OCR software information switching correction unit 32-2 executes correction processes corresponding to specific OCR software, such as scaling, compression, and correction using invisible information.

[0113] As described above, according to this embodiment, the functions created by the common correction unit 32-1 can be standardized, thereby reducing the number of design steps and the number of maintenance steps.

[0114] (Sixth embodiment) Next, a sixth embodiment will be described.

[0115] The sixth embodiment differs from the first embodiment in that OCR software information is set according to the AI ​​determination. In the following explanation of the sixth embodiment, the explanation of the same parts as in the first embodiment will be omitted, and only the parts that are different from the sixth embodiment will be explained.

[0116] Here, Fig. 21 is a diagram illustrating automatic linking of image correction by AI according to the sixth embodiment. As shown in Fig. 21, the OCR software information setting unit 31 sets OCR software information according to the judgment of the AI ​​that has learned training data linking the document to be read (English document, invoice, medical record) with the success rate of OCR processing. The image correction unit 32 automatically determines the optimal image correction for the document to be read according to the OCR software information.

[0117] In this way, according to this embodiment, the user's time and effort for setting can be eliminated.

[0118] (Seventh embodiment) The reading device shown in the first embodiment may be installed in an image forming apparatus.

[0119] Fig. 22 is a diagram showing an example of the configuration of an image forming apparatus according to the eighth embodiment. Fig. 22 shows an example of an image forming apparatus 3, which is generally called a multifunction peripheral (MFP). The image forming apparatus 3 shown in Fig. 22 has a reading device (reading device main body 10 and ADF 20) on the top. The configuration of the reading device is a repetition of the description of the first embodiment, so a detailed description will be omitted here.

[0120] The image forming apparatus 3 shown in FIG. 22 has an image forming section 80 and a paper feeding section 90 below the reading device main body 10.

[0121] The image forming unit 80 prints the image read by the reading device main body 10 onto a recording paper, which is an example of a recording medium. The read image is a visible image or a near-infrared image.

[0122] The image forming section 80 includes an optical writing device 81, tandem imaging units (Y, M, C, K) 82, an intermediate transfer belt 83, and a secondary transfer belt 84. In the image forming section 80, the optical writing device 81 writes an image to be printed onto the photosensitive drum 820 of the imaging unit 82, and the toner image of each plate is transferred from each photosensitive drum 820 onto the intermediate transfer belt 83. The K plate is formed with K toner containing carbon black.

[0123] The imaging unit (Y, M, C, K) 82 has four rotatable photosensitive drums (Y, M, C, K) 820, and is provided with imaging elements including a charging roller, a developing unit, a primary transfer roller, a cleaner unit, and a static eliminator around each photosensitive drum 820. As each imaging element operates in a predetermined image creation process around each photosensitive drum 820, an image is formed on each photosensitive drum 820, and the image formed on each photosensitive drum 820 is transferred as a toner image onto the intermediate transfer belt 83 by a primary transfer roller.

[0124] Intermediate transfer belt 83 is stretched across a drive roller and a driven roller in the nip between each photosensitive drum 820 and each primary transfer roller. The toner image that has been primarily transferred onto intermediate transfer belt 83 is secondarily transferred onto recording paper on secondary transfer belt 84 by a secondary transfer device as intermediate transfer belt 83 travels. The recording paper is then transported to fixing device 85 as secondary transfer belt 84 travels, where the toner image is fixed onto the recording paper. The recording paper is then ejected to a paper ejection tray outside the machine.

[0125] For example, a paper feed unit 90 feeds out a desired recording paper from paper feed cassettes 91 and 92 that store recording paper of different sizes, and transports it by a transport means 93 consisting of various rollers to supply it to the secondary transfer belt 84.

[0126] Note that image forming unit 80 is not limited to one that forms images by the electrophotographic method as described above, but may also be one that forms images by an inkjet method. Furthermore, the image forming device is not limited to an MFP having at least two of the functions of a copy machine, a printer, a scanner, and a facsimile machine, but may be any image forming device such as a copier, a scanner, or a facsimile machine. The image forming device may also be, for example, a printer that receives image data generated by a separate image processing device via communication and prints the received image data.

[0127] Although the embodiments of the present invention have been described above, each embodiment is presented as an example and is not intended to limit the scope of the invention. These novel embodiments can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. Each of these embodiments is included within the scope and spirit of the invention, and is also included in the inventions described in the claims and their equivalents.

[0128] For example, aspects of the present invention are as follows. <1> an image capturing unit that captures an image; an OCR software information setting unit that sets OCR software information according to the OCR software in the subsequent stage; an image correction unit that performs image correction on the image captured by the image capturing unit according to the OCR software information set in the OCR software information setting unit; an image data output unit that outputs the image corrected by the image correction unit; An image processing device comprising: <2> the image correction unit executes a process of lightening chromatic colors when the set OCR software information is OCR software without color removal or seal impression removal. Characterized by <1> The image processing device according to claim 1. <3> When the set OCR software information is OCR software without color removal or seal impression removal, the image correction unit executes a process of correcting red to be lighter. Characterized by <1> or <2> The image processing device according to claim 1. <4> the image correction unit executes a process of switching the compression level in accordance with the upper limit setting of the file size of the OCR software corresponding to the set OCR software information. Characterized by <1> Or <3> 10. The image processing device according to claim 9, wherein: <5> the image correction unit performs a process of turning off a subsampling process that thins out data, and performs compression using a low-compression table; Characterized by <4> The image processing device according to claim 1. <6> the image correction unit turns off subsampling, which thins out data, and performs compression using a normal table; Characterized by <4> The image processing device according to claim 1. <7> the image correction unit turns on a subsampling process for thinning out data, and performs compression using a low-compression table; Characterized by <4> The image processing device according to claim 1. <8> the image correction unit or the image capturing unit executes a process of switching resolution or image correction in accordance with an upper limit of a file size of the OCR software corresponding to the set OCR software information. Characterized by <1> Or <7> 10. The image processing device according to claim 9, wherein: <9> the image capturing unit captures an image by switching the resolution in accordance with the OCR software information set in the OCR software information setting unit; Characterized by <8> The image processing device according to claim 1. <10> The image correction unit executes a process of switching the resolution in accordance with the OCR software information set in the OCR software information setting unit. Characterized by <8> The image processing device according to claim 1. <11> the image capturing unit includes a visible image capturing unit for capturing a visible image and an invisible image capturing unit for capturing an invisible image, the image correction unit corrects the visible image captured by the visible image capturing unit by using the invisible image; Characterized by <1> The image processing device according to claim 1. <12> the image capturing unit includes a visible image capturing unit for capturing a visible image and an invisible image capturing unit for capturing an invisible image, at least one of the visible image capturing unit and the invisible image capturing unit switches its operation in accordance with the OCR software information set in the OCR software information setting unit; Characterized by <1> The image processing device according to claim 1. <13> The image correction unit a common correction unit that executes a common correction process for various OCR software; an OCR software information switching correction unit that executes correction processing corresponding to a specific OCR software; Equipped with Characterized by <1> Or <12> 10. The image processing device according to claim 9, wherein: <14> The image correction unit turns off image correction when the set OCR software information is OCR software that uses AI. Characterized by <1> Or <13> 10. The image processing device according to claim 9, wherein: <15> The image correction unit refers to data in which OCR software information, the presence or absence of AI, color removal / imprint removal function, and image correction content that defines what kind of image correction to perform are registered, and performs image correction that is set according to the OCR software information set in the OCR software information setting unit. Characterized by <1> Or <14> 10. The image processing device according to claim 9, wherein: <16> The OCR software information setting unit sets the OCR software information according to the selected OCR software name. Characterized by <1> Or <15> 10. The image processing device according to claim 9, wherein: <17> The OCR software information setting unit sets the OCR software information according to the function of the selected OCR software. Characterized by <1> Or <16> 10. The image processing device according to claim 9, wherein: <18> The OCR software information setting unit sets the OCR software information according to the determination of AI. Characterized by <1> Or <17> 10. The image processing device according to claim 9, wherein: <19> The OCR software information setting unit sets the OCR software information in accordance with an instruction from the OCR software in a subsequent stage. Characterized by <1> Or <18> 10. The image processing device according to claim 9, wherein: <20> <1> Or <19> an image processing device according to any one of the above; an image forming unit that forms an image based on image data generated by the image processing device; An image forming apparatus comprising: <21> a computer that controls an image processing device that includes an image capturing unit that captures an image; an OCR software information setting unit that sets OCR software information according to the OCR software in the subsequent stage; an image correction unit that performs image correction on the image captured by the image capturing unit according to the OCR software information set in the OCR software information setting unit; an image data output unit that outputs the image corrected by the image correction unit; A program to function as a <22> An image correction method in an image processing device including an image capturing unit that captures an image, an OCR software information setting step for setting OCR software information according to the OCR software in the subsequent stage; an image correction step of performing image correction set in accordance with the OCR software information set in the OCR software information setting step on the image captured by the image capturing unit; an image output step of outputting the image corrected by the image correction step; An image correction method comprising: [Explanation of symbols]

[0129] 1. Image processing device 3. Image forming device 31 OCR software information setting section 32 Image correction section 32-1 Common correction section 32-2 OCR software information switching correction section 33 Image data output unit 40 Image capturing unit 40-1 Visible image capturing unit 40-2 Invisible image capturing unit 80 Image forming unit [Prior art documents] [Patent documents]

[0130] [Patent Document 1] Japanese Patent Publication No. 2022-012252

Claims

1. an image capturing unit that captures an image; an OCR software information setting unit that sets OCR software information according to the OCR software in the subsequent stage; an image correction unit that performs image correction on the image captured by the image capturing unit according to the OCR software information set in the OCR software information setting unit; an image data output unit that outputs the image corrected by the image correction unit; An image processing device comprising:

2. the image correction unit executes a process of lightening chromatic colors when the set OCR software information is OCR software without color removal or seal impression removal.

2. The image processing device according to claim 1, wherein:

3. the image correction unit executes a process of correcting red to be lighter when the set OCR software information is OCR software without color removal or seal impression removal.

2. The image processing device according to claim 1, wherein:

4. the image correction unit executes a process of switching a compression level in accordance with a setting of an upper limit of a file size of the OCR software corresponding to the set OCR software information.

2. The image processing device according to claim 1, wherein:

5. the image correction unit performs a process of turning off a subsampling process that thins out data, and performs compression using a low-compression table; 5. The image processing device according to claim 4.

6. the image correction unit turns off subsampling, which thins out data, and performs compression using a normal table; 5. The image processing device according to claim 4.

7. the image correction unit turns on a subsampling process for thinning out data, and performs compression using a low-compression table; 5. The image processing device according to claim 4.

8. the image correction unit or the image capturing unit executes a process of switching resolution or image correction in accordance with an upper limit of a file size of the OCR software corresponding to the set OCR software information.

2. The image processing device according to claim 1, wherein:

9. the image capturing unit captures an image by switching a resolution in accordance with the OCR software information set in the OCR software information setting unit; 9. The image processing device according to claim 8,

10. The image correction unit executes a process of switching the resolution in accordance with the OCR software information set in the OCR software information setting unit.

9. The image processing device according to claim 8,

11. the image capturing unit includes a visible image capturing unit for capturing a visible image and an invisible image capturing unit for capturing an invisible image, the image correction unit corrects the visible image captured by the visible image capturing unit by using the invisible image; 2. The image processing device according to claim 1, wherein:

12. the image capturing unit includes a visible image capturing unit for capturing a visible image and an invisible image capturing unit for capturing an invisible image, at least one of the visible image capturing unit and the invisible image capturing unit switches its operation in accordance with the OCR software information set in the OCR software information setting unit; 2. The image processing device according to claim 1, wherein:

13. The image correction unit a common correction unit that executes a common correction process for various types of OCR software; an OCR software information switching correction unit that executes correction processing corresponding to a specific OCR software; Equipped with 2. The image processing device according to claim 1, wherein:

14. The image correction unit turns off image correction when the set OCR software information is OCR software using AI.

2. The image processing device according to claim 1, wherein:

15. The image correction unit refers to data in which OCR software information, the presence or absence of AI, color removal / imprint removal function, and image correction content that defines what kind of image correction to perform are registered, and performs image correction that is set according to the OCR software information set in the OCR software information setting unit.

2. The image processing device according to claim 1, wherein:

16. the OCR software information setting unit sets the OCR software information in accordance with the selected OCR software name; 2. The image processing device according to claim 1, wherein:

17. the OCR software information setting unit sets the OCR software information in accordance with the function of the selected OCR software; 2. The image processing device according to claim 1, wherein:

18. The OCR software information setting unit sets the OCR software information according to the determination of the AI.

2. The image processing device according to claim 1, wherein:

19. The OCR software information setting unit sets the OCR software information in response to an instruction from the OCR software in a subsequent stage.

2. The image processing device according to claim 1, wherein:

20. An image processing device according to any one of claims 1 to 19; an image forming unit that forms an image based on image data generated by the image processing device; An image forming apparatus comprising:

21. a computer that controls an image processing device that includes an image capturing unit that captures an image; an OCR software information setting unit that sets OCR software information corresponding to the OCR software in the subsequent stage; an image correction unit that performs image correction on the image captured by the image capturing unit according to the OCR software information set in the OCR software information setting unit; an image data output unit that outputs the image corrected by the image correction unit; A program to function as a

22. An image correction method in an image processing device including an image capturing unit that captures an image, an OCR software information setting step for setting OCR software information corresponding to the OCR software in the subsequent stage; an image correction step of performing image correction set in accordance with the OCR software information set in the OCR software information setting step on the image captured by the image capturing unit; an image output step of outputting the image corrected by the image correction step; An image correction method comprising:

Images