Image correction device, image correction method, and image forming device

Abstract

To provide an image forming device capable of correcting a photo image deteriorated over time according to an elapsed period even when a photographed date is not photographed in the photo image.SOLUTION: An image forming device for correcting a photo image formed on a sheet includes: an image reader that captures the photo image; a network communication circuit 156 that obtains the formation period of the photo image estimated by artificial intelligence; an input / output circuit 160 that obtains a correction coefficient for correcting the deterioration of the photo image corresponding to an elapsed period from the obtained formation period; and an image processing circuit 155 that corrects the captured photo image based on the obtained correction coefficient.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a technique for correcting a photographic image formed on a sheet. [Background technology]

[0002] Services have been offered to convert photographic images captured on silver halide negative film into image data. Because photographic images are formed on silver halide negative film through chemical reactions, storing them for a long period of time can lead to some deterioration in image quality, such as fading. Therefore, converting photographic images from silver halide negative film that have been stored for a long period of time directly into digital data is undesirable, as the image data will be stored in a state of degraded image quality.

[0003] According to Patent Document 1, statistical data indicating the deterioration characteristics of the image quality of a photographic image is collected in advance according to the period during which the silver halide negative film has been stored. If the shooting date is imprinted in the lower right corner of the photographic image on the silver halide negative film using the date function of the silver halide camera, the shooting date in the lower right corner of the photographic image is read using an OCR function with a scanner. When the photographic image is digitized, correction processing of the photographic image is performed based on the statistical data indicating the deterioration characteristics of the image quality of the photographic image according to the period of time elapsed since the shooting date read by the OCR function. In this way, the photographic image can be corrected according to the degree of image quality deterioration during the storage period. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2001-331781 Summary of the Invention [Problem to be solved by the invention]

[0005] As described above, according to Patent Document 1, the photographic date recorded in the photographic image is read using an OCR function, and the photographic image is corrected according to the degree of deterioration in the image quality of the photographic image over the period of time that has elapsed since the read photographic date.

[0006] However, since not all photographic images on silver halide negative film have the date of shooting imprinted on them, it is sometimes impossible to know how much time has passed since the photograph was taken. In this case, there is a problem that the quality of deteriorated photographic images cannot be appropriately corrected according to the amount of time that has passed. The same problem occurs when photographic images are printed on photographic paper.

[0007] The present disclosure aims to solve the above problem and provide an image correction device, an image correction method, and an image forming device that can correct photographic images that have deteriorated over time according to the amount of time that has passed, even if the date of shooting is not recorded in the photographic image. [Means for solving the problem]

[0008] In order to achieve the above object, one aspect of the present disclosure is an image correction device that corrects a photographic image formed on a sheet, the image correction device including: an image acquisition means that acquires a photographic image; a time acquisition means that acquires a formation time of the photographic image estimated by artificial intelligence; a coefficient acquisition means that acquires a correction coefficient for correcting deterioration of the photographic image according to the elapsed time since the acquired formation time; and an image correction means that corrects the acquired photographic image based on the acquired correction coefficient. The artificial intelligence uses at least information about the shape of an object in the photographic image to infer the time when the photographic image was created. It is characterized by:

[0009] Here, the system may further include a recognition means having artificial intelligence and using the artificial intelligence to infer the time when the photographic image was created as a feature of an object in the acquired photographic image, and the time acquisition means may acquire the time when the photographic image was created from the recognition means.

[0010] Here, the image correction device is connected to a server device via a network, and further includes a transmission means for transmitting the photographic image to the server device, and the server device has artificial intelligence and a recognition means for inferring the time of creation of the photographic image as a feature of an object in the received photographic image using the artificial intelligence, and the time acquisition means may acquire the time of creation from the server device.

[0011] Here, the image processing device further comprises a storage means for storing a correction coefficient for correcting deterioration of a photographic image in association with the elapsed time from the acquired creation time, coefficient The acquisition means may acquire the correction coefficient by reading from the storage means the correction coefficient corresponding to the period of time that has elapsed since the acquired formation time.

[0012] Here, the recognition means may further extract objects and object types within the photographic image by inference, the storage means may store the correction coefficients for each object type in association with the elapsed period, the coefficient acquisition means may read out from the storage means the correction coefficients corresponding to the extracted object type and the elapsed period since its formation, and the image correction means may correct the extracted object based on the read correction coefficients.

[0013] Here, the system may further include a storage means for storing a correction coefficient for correcting deterioration of a photographic image, for each paper type of the sheet, in association with the period of time that has elapsed since the acquired formation time, and a paper type acquisition means for acquiring the paper type of the sheet, and the coefficient acquisition means may acquire the correction coefficient from the storage means by reading out the correction coefficient that corresponds to the period of time that has elapsed since the acquired formation time, depending on the acquired paper type.

[0014] Furthermore, one aspect of the present disclosure is an image forming apparatus equipped with the above-described image correction device, the image forming apparatus including a method information acquisition means for acquiring method information indicating an image formation method used in the image forming apparatus, the image correction device further including a memory means for storing, for each image formation method, a correction coefficient for correcting deterioration of a photographic image in correspondence with the period of time elapsed since the acquired formation time, the coefficient acquisition means may acquire the correction coefficient corresponding to the period of time elapsed since the acquired formation time from the memory means in accordance with the acquired method information.

[0015] Another aspect of the present disclosure may be an image forming apparatus including the image correction device described above.

[0016] Another aspect of the present disclosure is an image correction method used in an image correction device that corrects a photographic image formed on a sheet, the image correction method comprising: an image acquisition step of acquiring a photographic image; using at least information on the shape of the object in the photographic image The time of creation of the inferred photographic image 、 The method is characterized by including an acquisition time acquisition step, a coefficient acquisition step for acquiring a correction coefficient for correcting deterioration of the photographic image according to the time elapsed since the acquired formation time, and an image correction step for correcting the acquired photographic image based on the acquired correction coefficient. [Effects of the Invention]

[0017] According to the aspects of the present disclosure, even if the photographic image does not include the date of shooting, it is possible to correct photographic images that have deteriorated over time using a correction coefficient according to the period of time that has elapsed since the time of creation, which is inferred from the photographic image, thereby achieving the excellent effect of correcting the photographic image. [Brief explanation of the drawings]

[0018] [Figure 1] 1 shows the configuration of an image correction system 1 according to a first embodiment. [Figure 2] 10(a) is a block diagram showing the configuration of the control circuit 100 of the image forming apparatus 5. FIG. 10(b) shows an example of the data structure of the correction coefficient table 171. FIG. [Figure 3](a) to (c) show photographic images 201, 211, and 221 obtained by scanning. [Figure 4] FIG. 2 is a block diagram showing the configuration of a server device 3. [Figure 5] FIG. 3 is a block diagram showing the configuration of a neural network 350. [Figure 6] 1A is a schematic diagram showing one neuron U of the neural network 350. FIG. 1B is a neuron setting table 360 ​​set in the neural network 350. FIG. [Figure 7] 1A is a diagram schematically illustrating a data propagation model during pre-learning (training) in a neural network 350. FIG. 1B is a diagram schematically illustrating a data propagation model during practical inference in a neural network 350. [Figure 8] 3 is a flowchart showing the operation of the image correction system 1. [Figure 9] (a) shows an example of the data structure of the correction coefficient table 241. (b) shows an example of the data structure of the correction coefficient table 251. (c) shows an example of the data structure of the correction coefficient table 261. [Figure 10] 10A shows an example of the data structure of the correction coefficient table 271. FIG. 10B shows an example of the data structure of the correction coefficient table 281. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0019] 1. First Embodiment 1.1 Image Correction System 1 An image correction system 1 according to the first embodiment will be described with reference to FIG.

[0020] The image correction system 1 is composed of a server device 3 and an image forming device 5 , and the server device 3 and the image forming device 5 are connected via a network 2 .

[0021] The image forming device 5 reads a photographic image that has faded over time from a photographic original. If the photographic image has the date of initial image formation printed on it, the image forming device 5 extracts that date from the photographic image as the time of formation. If the date of initial image formation is not printed on the photographic image, the image forming device 5 transmits the photographic image to the server device 3.

[0022] The server device 3 receives a photographic image from the image forming device 5, and estimates the time when the photographic image was formed from the photographic image using a neural network 350 (described later), and recognizes and extracts objects and object types contained in the photographic image. The server device 3 transmits to the image forming device 5 the time when the photographic image was formed, and the extracted objects and object types.

[0023] The image forming device 5 (image correction device) corrects the objects in the photographic image using a correction coefficient according to the time when the photographic image was formed.

[0024] By this correction, the image forming device 5 can make a photographic image that has faded over time closer to the original photographic image before fading.

[0025] 1.2 Image forming device 5 (1) Configuration of Image Forming Apparatus 5 As shown in FIG. 1, the image forming apparatus 5 is a tandem color multifunction peripheral (MFP) having the functions of a scanner, a printer, and a copier.

[0026] As shown in this figure, the image forming device 5 has a paper feed unit 13 at the bottom of the housing that stores and feeds recording sheets. Above the paper feed unit 13, a printer 12 that forms images by electrophotography is provided. Further above the printer 12, an image reader 11 that reads a document and generates image data, and an operation panel 19 that displays an operation screen and accepts input operations from the user are provided.

[0027] The image reader 11 has an automatic document feeder. The automatic document feeder feeds photographic documents, etc., set in a document tray, one by one to a document glass plate via a feed path. The image reader 11 (image acquisition means) reads photographic documents, etc., fed to a predetermined position on the document glass plate by the automatic document feeder, or photographic documents, etc., placed on the document glass plate by a user, by moving a scanner, and obtains image data consisting of multi-value digital signals of red (R), green (G), and blue (B). Here, the image reader 11 may read images on photographic paper or silver halide film.

[0028] The image data for each color component obtained by the image reader 11 undergoes various data processing in the control circuit 100. When the image data is printed, the control circuit 100 further converts it into image data for each reproduction color: yellow (Y), magenta (M), cyan (C), and black (K).

[0029] The printer 12 comprises an intermediate transfer belt 21 stretched between a drive roller, a driven roller, and a backup roller, a secondary transfer roller 22, image forming units 20Y, 20M, 20C, and 20K arranged at predetermined intervals along the running direction X of the intermediate transfer belt 21 facing the intermediate transfer belt 21, a fixing unit 50, a control circuit 100, and the like.

[0030] Imaging units 20Y, 20M, 20C, and 20K respectively form toner images of Y, M, C, and K colors. As an example, imaging unit 20K includes a photosensitive drum as an image carrier, an LED array for exposing and scanning the surface of the photosensitive drum, a charging device, a developing unit with a developing roller, a cleaner, and a primary transfer roller. Imaging units 20Y, 20M, and 20C have the same configuration as imaging unit 20K.

[0031] The paper feed section 13 is made up of paper feed cassettes 60, 61, and 62 that accommodate recording sheets of different sizes, and pickup rollers 63, 64, and 65 that feed the recording sheets from each paper feed cassette onto a conveying path.

[0032] In the image forming unit 20K, the peripheral surface of the photosensitive drum is uniformly charged to a negative polarity by a charging device and exposed to light by an LED array, forming an electrostatic latent image on the surface of the photosensitive drum. The electrostatic latent image is developed by a developing device, forming a K toner image on the surface of the photosensitive drum.

[0033] The toner images are transferred sequentially onto the surface of the intermediate transfer belt 21 by the electrostatic action of a primary transfer roller disposed on the back side of the intermediate transfer belt 21. The same is true for each of the image forming units 20Y to 20C.

[0034] The image formation timing of each color is shifted so that the toner images of Y, B, and K colors are transferred onto the intermediate transfer belt 21 in a multi-layered manner.

[0035] On the other hand, a recording sheet is fed from one of the paper feed cassettes in the paper feed unit 13 in accordance with the image forming operation by the image forming units 20Y to 20K.

[0036] The recording sheet is transported on the transport path to a secondary transfer position where a secondary transfer roller 22 and a backup roller face each other with the intermediate transfer belt 21 sandwiched therebetween, and at the secondary transfer position, the Y to K color toner images that have been multiply transferred on the intermediate transfer belt 21 are secondarily transferred onto the recording sheet by the electrostatic action of the secondary transfer roller 22. The recording sheet onto which the Y to K color toner images have been secondarily transferred is further transported to the fixing unit 50.

[0037] The toner image on the surface of the recording sheet is fused and fixed to the surface of the recording sheet by heat and pressure as it passes through the fixing nip formed between the heating roller of the fixing section 50 and the pressure roller pressed against it, and after passing through the fixing section 50, the recording sheet is sent to the discharge tray 15.

[0038] The operation panel 19 is provided with a display unit consisting of an LCD panel or the like, which displays settings made by the user, various messages, etc. The operation panel 19 (paper type acquisition means) accepts operation instructions from the user. The operation instructions include an instruction to start copying, setting the number of copies, setting copy conditions, setting the data output destination, a scan instruction to scan a photographic original, a correction instruction to correct color fading in a photographic image, and specification of the paper type of the original to be scanned (plain paper, photographic paper, etc.). The operation panel 19 notifies the control circuit 100 of the operation instructions.

[0039] (2) Control circuit 100 As shown in FIG. 2(a), the control circuit 100 is composed of a CPU (Central Processing Unit) 151, a ROM (Read Only Memory) 152, a RAM (Random Access Memory) 153, an image memory 154, an image processing circuit 155, a network communication circuit 156, a scanner control circuit 157, an input / output circuit 158, a printer control circuit 159, an input / output circuit 160, a memory circuit 161, a bus 166, etc.

[0040] The CPU 151, ROM 152, RAM 153, image memory 154, image processing circuit 155, network communication circuit 156, scanner control circuit 157, input / output circuit 158, printer control circuit 159, and input / output circuit 160 are connected to one another via a bus 166. The memory circuit 161 is connected to the input / output circuit 160.

[0041] The RAM 153 temporarily stores various control variables and data such as the number of copies set on the operation panel 19, and also provides a work area when the CPU 151 executes a program.

[0042] The ROM 152 stores control programs for executing various jobs such as copy operations, etc. The ROM 152 also stores printing method information indicating the printing method (electrophotographic method in this example) of the image forming apparatus 5.

[0043] The CPU 151 operates according to a control program stored in the ROM 152 .

[0044] The image memory 154 temporarily stores images read by the image reader 11, images included in a print job, and the like.

[0045] The image processing circuit 155 receives correction coefficients from the integrated control unit 142, which will be described later. Next, the image processing circuit 155 (image correction means) corrects the photographic image by adding the received correction coefficients to the gradation values ​​of each pixel for all pixels in the photographic image stored in the image memory 154, or for a specified object in the photographic image, using the following formula. Note that an example in RGB space is shown here.

[0046] Red pixel gradation value = Red pixel gradation value + Red correction coefficient Green pixel tone value = Green pixel tone value + Green correction coefficient Blue pixel tone value = Blue pixel tone value + Blue correction coefficient In addition, the image processing circuit 155 performs various data processing on the image data of each color component of R, G, and B obtained by the image reader 11, for example, when the image data is printed, and converts it into image data of each reproduction color of Y, M, C, and K.

[0047] The network communication circuit 156 transmits data to an external device via the network 2. For example, the network communication circuit 156 (transmitting means) transmits a photographic image to the server device 3 via the network 2. The network communication circuit 156 also receives data from an external device via the network 2. For example, the network communication circuit 156 (time acquiring means) receives, from the server device 3, the time when the photographic image was formed, the object, and the type of object via the network 2.

[0048] Scanner control circuit 157 controls image reader 11 to execute the operation of reading a document. Scanner control circuit 157 writes the image obtained by reading with image reader 11 into image memory 154. For example, when a photographic image printed on photographic paper is read by image reader 11, scanner control circuit 157 writes the obtained photographic image into image memory 154 as photographic image 181.

[0049] The input / output circuit 158 ​​receives an input signal from the operation panel 19 and outputs the received input signal to the main control unit 141. It also receives an image from the main control unit 141 and outputs the received image to the operation panel 19 for display.

[0050] The printer control circuit 159 controls the printer 12 to perform image forming operations.

[0051] The input / output circuit 160 writes data to the memory circuit 161 under the control of the integrated control unit 142, and reads data from the memory circuit 161. As an example, the input / output circuit 160 (coefficient acquisition means) reads out a correction coefficient corresponding to the formation time from the correction coefficient table 171 in the memory circuit 161 under the control of the image processing circuit 155, and outputs the read correction coefficient to the image processing circuit 155.

[0052] The memory circuit 161 will be described next.

[0053] (3) Memory circuit 161 The storage circuit 161 is, for example, configured from a semiconductor memory.

[0054] The memory circuit 161 (storage means) stores in advance a correction coefficient table 171 for correcting a person's face image. In addition, the memory circuit 161 stores in advance other correction coefficient tables according to the application, for example, according to the type of object included in the photographic image, according to the paper type of the document to be scanned, or according to the printing method in the image forming apparatus 5. Here, the type of object indicates, for example, whether the object is the sea, the sky, a forest, a person's face, etc., and the printing method is, for example, an electrophotographic method, an inkjet method, etc.

[0055] (Correction coefficient table 171) The correction coefficient table 171 stores correction coefficients for correcting facial images of people in photographic images that have faded over time.

[0056] The correction coefficient table 171 is generated by collecting statistical data in advance that indicates the deterioration characteristics of the image quality of photographic images formed on photographic paper according to the period for which the photographic paper on which the photographic images are formed has been stored.

[0057] 2B, the correction coefficient table 171 includes a plurality of correction coefficient information 172. Each correction coefficient information 172 is made up of a set of a year range 173 and a correction coefficient 174.

[0058] The year range 173 indicates the year the photographic image was formed on the photographic paper.

[0059] As an example, as shown in FIG. 2(b), each correction coefficient information 172 includes, as year ranges 173, "up to 2020," "up to 2015," "up to 2010," "up to 2005," "up to 2000," and "up to 1950."

[0060] Here, the year range 173 "~1950" indicates the period from 1950 to 1999, the year range 173 "~2000" indicates the period from 2000 to 2004, the year range 173 "~2005" indicates the period from 2005 to 2009, the year range 173 "~2010" indicates the period from 2010 to 2014, the year range 173 "~2015" indicates the period from 2015 to 2019, and the year range 173 "~2020" indicates the period from 2020 onwards.

[0061] If the formation date is before 1945, it may be included in the year range 173 "~1950".

[0062] The correction coefficients 174 are used when correcting a person's face image among the photographic images formed within the year range 173 corresponding to the correction coefficient 174. The correction coefficients 174 consist of an R (red) correction coefficient 174a, a G (green) correction coefficient 174b, and a B (blue) correction coefficient 174c.

[0063] The R correction coefficient 174a indicates a correction coefficient applied to red pixels, the G correction coefficient 174b indicates a correction coefficient applied to green pixels, and the B correction coefficient 174c indicates a correction coefficient applied to blue pixels.

[0064] When correcting each pixel of a photographic image, the value of the R correction coefficient 174a, the value of the G correction coefficient 174b, and the value of the B correction coefficient 174c are added to the R component gradation value, the G component gradation value, and the B component gradation value of each pixel, respectively.

[0065] This correction makes it possible to make the facial image of a person in a photographic image that has faded over time closer to the original facial image of the person before the fade.

[0066] Other correction coefficient tables will be described later.

[0067] (4) Example of image data 3(a) to 3(c) show, as an example, photographic images 201, 211, and 221 obtained by reading with the image reader 11. In FIG.

[0068] Photo image 201 shows a school graduation sign along with several people.

[0069] Additionally, photo image 211 also shows a sign directing people to a sports event venue, along with multiple people. The people in photo image 211 are wearing clothes with designs that were popular at the time the photo was taken. Additionally, a car that was in use at the time may also be visible. The design of the clothes and the shape of the car are used as features for identifying the time the photo was taken.

[0070] Additionally, multiple people appear in the photographic image 211. The date the photograph was taken is printed in the lower right corner of the photographic image 211.

[0071] (5) Main control unit 141 The main control unit 141 is made up of a CPU 151, a ROM 152, and a RAM 153. The CPU 151 operates in accordance with a control program stored in the ROM 152, whereby the main control unit 141 performs its functions.

[0072] As shown in FIG. 2( a ), the main control unit 141 is made up of an overall control unit 142 and an OCR processing unit 143 .

[0073] (a) General control unit 142 The CPU 151 operates in accordance with the control program, and the overall control unit 142 comprehensively controls the image memory 154, image processing circuit 155, network communication circuit 156, scanner control circuit 157, input / output circuit 158, printer control circuit 159, input / output circuit 160, etc. The overall control unit 142 also controls the OCR processing unit 143.

[0074] The overall control unit 142 (method information acquisition means) reads out from the ROM 152 the printing method information indicating the printing method.

[0075] The general control unit 142 receives operation instructions from the operator via the input / output circuit 158 ​​from the operation panel 19. The operation instructions include a scan instruction to scan a photographic original, a correction instruction to correct color fading in a photographic image, and a specification of the paper type of the original to be scanned.

[0076] When a scan instruction is received, the integrated control unit 142 controls the scanner control circuit 157 to read the photographic original and generate a photographic image using the image reader 11. The obtained photographic image is written to the image memory 154 as a photographic image 181.

[0077] Furthermore, the overall control unit 142 determines whether or not a correction instruction has been received.

[0078] If it is determined that a correction instruction has not been received, the overall control unit 142 does not execute correction processing on the photographic image.

[0079] If it is determined that a correction instruction has been received, the overall control unit 142 controls the OCR processing unit 143 to output the photographic image and read date data from the photographic image.

[0080] The overall control unit 142 receives the processing result from the OCR processing unit 143. If the processing result indicates that date data exists in the photographic image, the overall control unit 142 receives the date data from the OCR processing unit 143 and identifies the received date data as the formation time.

[0081] If the processing result indicates that date data does not exist in the photographic image, the overall control unit 142 controls the network communication circuit 156 to transmit the obtained photographic image to the server device 3 via the network 2.

[0082] The central control unit 142 receives one or more objects, the types of those objects, and the creation times thereof from the server device 3 via the network 2 and the network communication circuit 156.

[0083] The integrated control unit 142 selects one correction coefficient table from the multiple correction coefficient tables stored in the memory circuit 161 for each object, using the object type, designated paper type, and printing method information.

[0084] When the integrated control unit 142 receives the formation time, it controls the input / output circuit 160 to read the selected correction coefficient table from the memory circuit 161. Next, the integrated control unit 142 extracts a correction coefficient corresponding to the received formation time from the read correction coefficient table.

[0085] Next, the general control unit 142 outputs the extracted correction coefficients to the image processing circuit 155.

[0086] Furthermore, the general control unit 142 designates the received object from the photograph image to the image processing circuit 155 .

[0087] The general control unit 142 controls the image processing circuit 155 so as to correct the received object in the photograph image.

[0088] The general control unit 142 receives a notification from the image processing circuit 155 that correction of the photographic image has been completed.

[0089] When receiving a notification from the image processing circuit 155 that the correction of the photographic image has been completed, the overall control unit 142 controls the printer control circuit 159 to print the corrected photographic image. As a result, the printer 12 prints the corrected photographic image.

[0090] (b) OCR processing unit 143 The OCR processing unit 143 receives the photographic image from the general control unit 142 .

[0091] OCR processing unit 143 performs OCR processing on the received photographic image. If the photographic image contains date data as a result, OCR processing unit 143 outputs a processing result indicating that date data exists and the obtained date data to overall control unit 142. On the other hand, if the photographic image does not contain date data, OCR processing unit 143 outputs a processing result indicating that date data does not exist to overall control unit 142.

[0092] 1.3 Selection of correction coefficient table The integrated control unit 142 selects one correction coefficient table from the multiple correction coefficient tables stored in the memory circuit 161 using the object type, paper type designation, and printing method information, as shown below.

[0093] (1) The memory circuit 161 may further store a correction coefficient table 241 corresponding to blue, a correction coefficient table 251 corresponding to red, and a correction coefficient table 261 corresponding to green, as shown in FIGS. 9(a) to 9(c).

[0094] The correction coefficient table 241, the correction coefficient table 251, and the correction coefficient table 261 each have the same data structure as the correction coefficient table 171, and therefore a description of the correction coefficient table 241, the correction coefficient table 251, and the correction coefficient table 261 will be omitted.

[0095] The correction coefficient table 241 corresponding to blue can be used, for example, for blue sky objects, the correction coefficient table 251 corresponding to red can be used, for example, for sunset sky objects, and the correction coefficient table 261 corresponding to green can be used, for example, for forest objects.

[0096] As described above, the central control unit 142 receives the object type from the server device 3.

[0097] When the type of object indicates that it is a human face object, the overall control unit 142 selects the correction coefficient table 171 shown in FIG. 2(a).

[0098] When the type of object indicates that it is a blue sky object, the central control unit 142 selects the correction coefficient table 241 shown in FIG. 9(a).

[0099] When the type of object indicates that it is a sunset sky object, the central control unit 142 selects the correction coefficient table 251 shown in FIG. 9(b).

[0100] If the type of object indicates that it is a forest object, the central control unit 142 selects the correction coefficient table 261 shown in FIG. 9(c).

[0101] Next, the integrated control unit 142 extracts a correction coefficient corresponding to the received formation time from the selected correction coefficient table, and outputs the extracted correction coefficient to the image processing circuit 155 .

[0102] The image processing circuit 155 receives the correction coefficients and uses the received correction coefficients to correct the objects in the photographic image.

[0103] This allows correction of a deteriorated photographic image according to the type of object.

[0104] (2) As described above, the operation panel 19 accepts a user's designation of the paper type of the document to be scanned. The paper type may be, for example, plain paper, photographic paper, etc. The integrated control unit 142 receives the designation of the paper type from the operation panel 19 via the input / output circuit 158.

[0105] Moreover, the general control unit 142 receives the formation time.

[0106] The storage circuitry 161 may further store a correction coefficient table 271 corresponding to paper types, as shown as an example in FIG. 10(a).

[0107] The correction coefficient table 271 stores correction coefficients for correcting when photographic images formed on photographic paper and plain paper have faded over time.

[0108] The correction coefficient table 271 is generated by collecting statistical data in advance that indicates the deterioration characteristics of the image quality of photographic images formed on photographic paper and plain paper, depending on the period for which the photographic paper and plain paper on which the photographic images are formed have been stored.

[0109] The correction coefficient table 271 includes a plurality of correction coefficient information 272. Each correction coefficient information 272 is made up of a set of a year range 273 and a correction coefficient 274. The correction coefficient 274 includes a lightness L (274a) of photographic paper and a lightness L (274b) of plain paper.

[0110] The year range 273 indicates the year in which the photographic image was formed on photographic paper or plain paper.

[0111] As an example, as shown in FIG. 10(a), the correction coefficient table 271 includes, as year ranges 273, "up to 2020," "up to 2015," "up to 2010," "up to 2005," "up to 2000," and "up to 1950."

[0112] These are the same as the year range 173 in the correction coefficient table 171 shown in FIG. 2(b), and therefore the explanation will be omitted.

[0113] The lightness L of the photographic paper (274a) is used when correcting a photographic image formed on the photographic paper, and the lightness L of the plain paper (274b) is used when correcting a photographic image formed on the photographic paper.

[0114] The lightness L of the photographic paper (274a) and the lightness L of the plain paper (274b) each indicate lightness in the Lab color space.

[0115] When the overall control unit 142 receives a designation of the type of photographic paper, it extracts the lightness L (274a) of the photographic paper corresponding to the received designation of the paper type and the received formation time from the correction coefficient table 271. On the other hand, when the overall control unit 142 receives a designation of the type of plain paper, it extracts the lightness L (274b) of the plain photographic paper corresponding to the received designation of the paper type and the received formation time from the correction coefficient table 271.

[0116] The integrated control unit 142 outputs the extracted lightness L to the image processing circuit 155.

[0117] The image processing circuit 155 receives the lightness L from the general control unit 142 and corrects the lightness L of the photographic image 181 by adding the received lightness L to the lightness L of the photographic image 181 .

[0118] (3) The image forming apparatus 5 performs printing by electrophotography. However, the image forming apparatus 5 may perform printing by inkjet printing. In this case, the ROM 152 of the image forming apparatus 5 stores printing method information indicating the inkjet printing method.

[0119] The storage circuitry 161 may further store a correction coefficient table 281 corresponding to the printing method, as shown as an example in FIG. 10(b).

[0120] Here, the correction coefficient table 281 corresponds to an electrophotographic image forming apparatus and an inexpensive inkjet image forming apparatus.

[0121] The correction coefficient table 281 stores correction coefficients for correcting fading of photographic images formed on photographic paper by electrophotographic image forming apparatuses and inexpensive inkjet image forming apparatuses over time.

[0122] The correction coefficient table 281 is generated by collecting statistical data in advance that indicates the deterioration characteristics of the image quality of photographic images formed on photographic paper according to the period for which the photographic paper on which the photographic images are formed has been stored.

[0123] The correction coefficient table 281 includes a plurality of correction coefficient information 282. Each correction coefficient information 282 is made up of a set of a year range 283 and a correction coefficient 284. The correction coefficient 284 includes a lightness L (284a) for inexpensive inkjet and a lightness L (284b) for electrophotography.

[0124] The year range 283 indicates the year in which the photographic image was formed on photographic paper or plain paper. The year range 283 is the same as the year range 173 in the correction coefficient table 171 shown in Figure 2(b), so a description thereof will be omitted.

[0125] The lightness L (284a) is used when correcting a photographic image formed on photographic paper by an inexpensive inkjet image forming device, and the lightness L (284b) is used when correcting a photographic image formed on photographic paper by an electrophotographic image forming device.

[0126] The lightness L (284a) and the lightness L (284b) each indicate lightness in the Lab color space.

[0127] The central control unit 142 reads out the printing method information from the ROM 152 .

[0128] If the printing method information read from the ROM 152 indicates the electrophotographic method, the integrated control unit 142 extracts from the correction coefficient table 281 the lightness L (284b) corresponding to the read printing method information and the received formation time.

[0129] On the other hand, if the printing method information read from the ROM 152 indicates the inkjet method, the central control unit 142 extracts from the correction coefficient table 281 the lightness L (284a) that corresponds to the read printing method information and the received formation time.

[0130] The integrated control unit 142 outputs the extracted lightness L to the image processing circuit 155.

[0131] The image processing circuit 155 receives the lightness L from the general control unit 142 and corrects the lightness L of the photographic image 181 by adding the received lightness L to the lightness L of the photographic image 181 .

[0132] This allows correction of deteriorated photographic images according to the printing method.

[0133] 1.4 Server device 3 (1) Configuration of server device 3 4, the server device 3 is composed of a CPU 301, a ROM 302, a RAM 303, a memory circuit 304, and a network communication circuit 311 connected to a bus B1, and a GPU (Graphics Processing Unit) 305, a ROM 306, a RAM 307, and a memory circuit 308 connected to a bus B2. The buses B1 and B2 are connected to each other.

[0134] The RAM 303 is made up of semiconductor memory and provides a work area when the CPU 301 executes a program.

[0135] The ROM 302 is also composed of a non-volatile semiconductor memory. The ROM 302 stores a control program for executing the functions of the server device 3. The control program may be stored in the storage circuit 304.

[0136] The CPU 301 operates according to a control program stored in the ROM 302 .

[0137] The network communication circuit 311 receives data from an external device via the network 2. Here, an example of the external device is the image forming device 5, and an example of the received data is a photographic image. Upon receiving the photographic image, the network communication circuit 311 writes the received photographic image as a photographic image 389 in the storage circuit 304. The network communication circuit 311 also transmits data to the external device via the network 2. Here, an example of the external device is the image forming device 5, and an example of the transmitted data is an object extracted by recognition from the photographic image 389, the type of the object, and the time of formation.

[0138] The memory circuit 304 is configured from a non-volatile semiconductor memory. Alternatively, the memory circuit 304 may be configured from a hard disk. The memory circuit 304 has an area for storing a photographic image 389 to be recognized.

[0139] The RAM 307 is made up of non-volatile semiconductor memory and provides a work area when the GPU 305 executes a program.

[0140] The ROM 306 stores a control program, which is a computer program for executing processing in a recognition processing unit 324 (described later).

[0141] The GPU 305 operates according to a control program stored in the ROM 306 .

[0142] The GPU 305 operates according to a control program stored in the ROM 306 using the RAM 307 as a work area, whereby the GPU 305 , the ROM 306 , and the RAM 307 constitute a recognition processing unit 324 .

[0143] The recognition processing unit 324 incorporates a neural network 350, which will be described later, as an artificial intelligence.

[0144] The neural network 350 incorporated in the recognition processing unit 324 performs its functions by the GPU 305 operating in accordance with a control program stored in the ROM 306 .

[0145] The recognition processing unit 324 (recognition means) performs recognition processing on a photographic image received from the main control unit 321 (described later) using a neural network 350 to infer objects in the photographic image, the object types, and the time when the photographic image was formed, and outputs the objects, object types, and the time when the photographic image was formed to the main control unit 321. Examples of objects include a person's entire body object, a person's face object, a vehicle object, a signboard object, a furniture object, an ocean object, a sky object, a planted object, a forest object, etc.

[0146] The memory circuitry 308 stores a neuron setting table 360, which will be described later.

[0147] (2) Neural Network 350 The neural network 350 will be described with reference to FIG.

[0148] (a) Structure of neural network 350 As shown in this figure, the neural network 350 is a hierarchical neural network having an input layer 350a, a feature extraction layer 350b, and a recognition layer 350c.

[0149] Here, a neural network is an information processing system that mimics the human neural network. In the neural network 350, the engineered neuron model, which corresponds to a nerve cell, is called a neuron U. The input layer 350a, the feature extraction layer 350b, and the recognition layer 350c each include multiple neurons U.

[0150] The input layer 350a typically consists of one layer. Each neuron U in the input layer 350a receives, for example, the pixel values ​​of each pixel constituting an image. The received image values ​​are output directly from each neuron U in the input layer 350a to the feature extraction layer 350b.

[0151] The feature extraction layer 350b extracts features from the data received from the input layer 350a (for example, all pixel values ​​constituting one image) and outputs them to the recognition layer 350c. This feature extraction layer 350b extracts the area occupied by the object from the received image through calculations in each neuron U.

[0152] The recognition layer 350c performs classification using the features extracted by the feature extraction layer 350b. The recognition layer 350c identifies the type and attributes of an object from the area of ​​the object extracted by the feature extraction layer 350b, for example, through calculations in each neuron U. Examples of object types include people, plants, furniture, vehicles, the sea, forests, and the sky. An example of an object attribute is the date on which the object was created (time of creation).

[0153] Neuron U is usually a multi-input, single-output element, as shown in Figure 6(a). Signals propagate in only one direction, and the input signal xi (i = 1, 2, ..., n) is multiplied by a certain neuron weight (SUwi) before being input to neuron U. This neuron weight represents the strength of the connection between neurons U arranged hierarchically. Neuron weights can be changed through learning. Neuron U outputs a value X, which is the sum of each input value (SUwi × xi) multiplied by the neuron weight SUwi minus a neuron threshold θU, after being transformed by a response function f(X). In other words, the output value y of neuron U is expressed by the following formula:

[0154] y=f(X) where: X=Σ(SUwi×xi)-θU As the response function, for example, a sigmoid function can be used.

[0155] Each neuron U in the input layer 350a typically does not have a sigmoid characteristic or a neuron threshold. Therefore, the input value is directly reflected in the output. On the other hand, each neuron U in the final layer (output layer) of the recognition layer 350c outputs the classification result of the recognition layer 350c.

[0156] The learning algorithm for the neural network 350 is, for example, backpropagation, which uses the steepest descent method to sequentially change the neuron weights of the recognition layer 350c and the neuron weights of the feature extraction layer 350b so that the squared error between the value (data) indicating the correct answer and the output value (data) from the recognition layer 350c is minimized.

[0157] (b) Training process The training process in neural network 350 will now be described.

[0158] The training step is a step of pre-learning the neural network 350. In the training step, pre-learning of the neural network 350 is performed using image data with correct answers (supervised and annotated) that has been obtained in advance.

[0159] Figure 7(a) shows a schematic diagram of the data propagation model during pre-learning.

[0160] Image data is input to input layer 350a of neural network 350 for each image, and output from input layer 350a to feature extraction layer 350b. Each neuron U in feature extraction layer 350b performs a computation with a neuron weight on the input data. Through this computation, feature extraction layer 350b extracts features (e.g., the area occupied by an object) from the input data, and outputs data indicating the extracted features to recognition layer 350c (step S51).

[0161] Each neuron U in the recognition layer 350c performs a computation with a neuron weight on the input data (step S52). This allows classification based on the above features (e.g., classification of object attributes). Data indicating the classification result is output from the recognition layer 350c.

[0162] The output value (data) of the recognition layer 350c is compared with a value indicating the correct answer, and the error (loss) between them is calculated (step S53). To reduce this error, the neuron weights of the recognition layer 350c and the neuron weights of the feature extraction layer 350b are sequentially changed (backpropagation) (step S54). This allows the recognition layer 350c and the feature extraction layer 350b to train.

[0163] The above training process results in the following neuron configuration table 360:

[0164] The training process described here is performed in a training device separate from the server device 3, and is not performed in the server device 3. The neuron setting table 360 ​​generated in the training device is transmitted to the server device 3, and the server device 3 receives the neuron setting table 360 ​​and writes the received neuron setting table 360 ​​into the memory circuit 308 (FIG. 4).

[0165] Here, examples of combinations of training data and values ​​indicating correct answers are as follows:

[0166] A photo of the vehicle and the date it was first sold A photo of the woman's clothing and the date it was taken A photo of a woman's hairstyle and the date it was taken A photo of the man's clothing and the date it was taken Photos of sporting events and the dates they were taken (c) Neuron setting table 360 As a result of the above training process, the neuron setting table 360 ​​shown in Fig. 6(b) is generated. As described above, the neuron setting table 360 ​​is stored in the memory circuit 308 shown in Fig. 4.

[0167] 6(b), the neuron setting table 360 ​​is made up of a plurality of pieces of neuron information 361. Each piece of neuron information 361 corresponds to each neuron U described above.

[0168] Each neuron information 361 includes a neuron number 362 , a neuron weight 363 , and a neuron threshold 364 .

[0169] The neuron number 362 is a number that identifies each neuron U.

[0170] Neuron weights 363 and neuron thresholds 364 are the neuron weights and neuron thresholds described above, respectively.

[0171] (d) On-site recognition process The actual recognition process in the neural network 350 will now be described.

[0172] FIG. 7(b) shows a data propagation model for when the neural network 350 learned by the above training process is used to perform actual recognition (e.g., identifying the area occupied by an object and recognizing the attributes of the object) using data obtained in the field as input.

[0173] In the on-the-spot recognition process in the neural network 350, feature extraction and recognition are performed using the trained feature extraction layer 350b and the trained recognition layer 350c (step S55).

[0174] (3) Main control unit 321 The main control unit 321 is made up of a CPU 301, a ROM 302, and a RAM 303. The CPU 301 operates in accordance with a control program stored in the ROM 302, whereby the main control unit 321 performs its functions.

[0175] The main control unit 321 controls the memory circuit 304, the network communication circuit 311, the recognition processing unit 324, and the like in an integrated manner.

[0176] When the photographic image 389 is written to the memory circuit 304, the main control unit 321 outputs the photographic image 389 written to the memory circuit 304 to the recognition processing unit 324 via the bus B1 and the bus B2.

[0177] Furthermore, the main control unit 321 receives the object, the object type, and the creation time from the recognition processing unit 324. The main control unit 321 controls the network communication circuit 311 to transmit the received object, object type, and creation time to the image forming device 5 via the network 2.

[0178] 1.5 Operation of Image Correction System 1 The operation of the image correction system 1 will be described with reference to the flowchart shown in FIG.

[0179] The central control unit 142 acquires the printing method by reading printing method information indicating the printing method from the ROM 152 (step S101).

[0180] Next, the central control unit 142 acquires the paper type of the document to be scanned from the operation panel 19 (step S102).

[0181] Next, when the integrated control unit 142 receives a scan instruction to scan a photographic original, it controls the scanner control circuit 157 to read the photographic original using the image reader 11 and generate a photographic image (step S103).

[0182] The central control unit 142 determines whether or not a correction instruction to correct color fading of the photographic image has been received (step S104). If it determines that a correction instruction has not been received ("No" in step S104), the central control unit 142 transfers control to step S112.

[0183] If it is determined that a correction instruction has been received ("Yes" in step S104), the overall control unit 142 controls the OCR processing unit 143 to output a photographic image and read date data from the photographic image. If date data exists in the image ("YES" in step S105), the overall control unit 142 identifies the read date data as the formation time (step S113). Next, the overall control unit 142 transfers control to step S109.

[0184] If date data does not exist in the image ("NO" in step S105), the overall control unit 142 controls the network communication circuit 156 to transmit the photographic image to the server device 3 via the network 2. The network communication circuit 156 transmits the photographic image to the server device 3 via the network 2 (step S106).

[0185] The network communication circuit 311 receives the photographic image from the image forming device 5 via the network 2 (step S106).

[0186] When the photographic image is received (step S106), the main control unit 321 outputs the received photographic image to the recognition processing unit 234 and instructs the recognition processing unit 234 to perform recognition processing on the received photographic image. The recognition processing unit 234 performs recognition processing on the received photographic image. As a result, the recognition processing unit 234 estimates and outputs the object, the object type, and the time of formation (step S107).

[0187] The main control unit 321 controls the network communication circuit 311 to transmit the obtained object, object type, and creation time to the image forming device 5 via the network 2. The network communication circuit 311 transmits the obtained object, object type, and creation time to the image forming device 5 via the network 2 (step S108).

[0188] The network communication circuit 156 receives the object, the type of object, and the creation time from the server device 3 via the network 2 (step S108).

[0189] The central control unit 142 selects a correction coefficient table according to the type of object, the printing method, the paper type of the document, etc. (step S109).

[0190] The integrated control unit 142 reads out the correction coefficient corresponding to the formation time from the correction coefficient table 171 in the storage circuit 161 via the input / output circuit 160, and outputs the read correction coefficient to the image processing circuit 155 (step S110).

[0191] The general control unit 142 controls the image processing circuit 155 to correct the photographic image, and the image processing circuit 155 corrects the photographic image using the read correction coefficients (step S111).

[0192] Next, the integrated control unit 142 controls the printer control circuit 159 to print the corrected photographic image. The printer 12 then prints the corrected photographic image (step S112).

[0193] This concludes the description of the operation of the image correction system 1.

[0194] 1.6 Summary As described above, the image forming device 5 corrects a photographic image that has faded over time using a correction coefficient according to the time when the photographic image was formed, which is obtained by inference using artificial intelligence. Through this correction, the image forming device 5 can make the faded photographic image closer to the original photographic image before fading.

[0195] 2 Other variations Although the present disclosure has been described based on the above-described embodiment, it goes without saying that the present disclosure is not limited to the above-described embodiment.

[0196] (1) The correction coefficient table 171 shown in Fig. 2(b) includes the year range in which the photographic image was formed and the correction coefficient, but is not limited to this.

[0197] The memory circuit 161 (memory means) may store a correction coefficient table that associates the correction coefficient with the time that has passed since the photographic image was formed.

[0198] The overall control unit 142 calculates the difference between the obtained formation time and the current time, searches the correction coefficient table for an elapsed period corresponding to the calculated difference, and reads out from the correction coefficient table a correction coefficient corresponding to the elapsed period obtained by the search. The overall control unit 142 outputs the read correction coefficient to the image processing circuit 155. The image processing circuit 155 receives the correction coefficient and corrects the image using the received correction coefficient.

[0199] (2) The image forming device 5 may be equipped with the same neural network as the neural network 350 of the server device 3, and may also be equipped with the same recognition processing unit as the recognition processing unit 324 of the server device 3. The recognition processing unit (recognition means) performs recognition processing on the photographic image received from the main control unit 321 using a neural network to infer the object, the object type, and the time when the photographic image was formed, and outputs the object, the object type, and the time when the photographic image was formed to the main control unit 321. The main control unit 321 (time acquisition means) receives the time when the photographic image was formed. The main control unit 321 also receives the object and the object type.

[0200] (3) The integrated control unit 142 of the image forming device 5 controls the printer control circuit 159 to print the corrected photographic image, and the printer 12 prints the corrected photographic image. However, this is not limited to this.

[0201] The overall control unit 142 controls the network communication circuit 156 to transmit the corrected photographic image to other terminal devices connected to the network 2, and the network communication circuit 156 may transmit the corrected photographic image to the other terminal devices.

[0202] (4) In the above embodiment, the image forming device 5 scans a photographic original to obtain a photographic image, and transmits the obtained photographic image to the server device 3. The server device 3 obtains the object, object type, and creation time from the photographic image by inference, and transmits the obtained object, object type, and creation time to the image forming device 5. The image forming device 5 selects a correction coefficient table using the object type, paper type of the original, and printing method information of the image forming device 5, and corrects the photographic image using the selected correction coefficient table. However, this is not limited to this. The following may also be done.

[0203] The image forming device scans a photographic original to obtain a photographic image, and transmits the obtained photographic image to the server device. The image forming device also transmits to the server device information on the paper type of the photographic image and the printing method of the image forming device.

[0204] The server device stores a correction coefficient table in advance depending on the application, similar to the image forming device 5. The server device also includes an image processing circuit similar to the image processing circuit 155 of the image forming device 5.

[0205] The server device obtains the object, the type of object, and the time of creation from the received photographic image by inference.

[0206] The server device selects a correction coefficient table using the type of object acquired by inference, the paper type of the received document, and the printing method information of the received image forming device. The image processing circuit of the server device (image correction device) corrects the photographic image using the selected correction coefficient table.

[0207] The server device transmits the corrected photographic image to the image forming device, which receives the corrected photographic image and prints the received photographic image.

[0208] (5) In the above embodiment, the image forming device 5 stores a correction coefficient table for each object type, document paper type, and printing method information. However, this is not limited to this, and the image forming device does not necessarily have to store a correction coefficient table.

[0209] A server device connected to the image forming apparatus via a network may store a correction coefficient table for each object type, document paper type, and printing method information.

[0210] The image forming device transmits the object type, document paper type, printing method information, and formation time to the server device. The server device receives the object type, document paper type, printing method information, and formation time, and selects from stored correction coefficient tables a correction coefficient table that corresponds to the received object type, document paper type, and printing method information. Next, the server device reads correction coefficient information that corresponds to the received formation time from the selected correction coefficient table, and transmits the read correction coefficient information to the image forming device.

[0211] The image forming device receives the correction coefficient information and corrects the photographic image using the received correction coefficient information.

[0212] (6) The above-described embodiments and modifications may be combined with each other. [Industrial Applicability]

[0213] The image correction device disclosed herein has the excellent effect of correcting a photographic image that has faded over time using a correction coefficient according to the time elapsed since the image was created, which is obtained by inferring from the photographic image, thereby making it closer to the original photographic image before fading, and is therefore useful as a technology for correcting photographic images. [Explanation of symbols]

[0214] 1. Image Correction System 2 Network 3 Server equipment 5. Image forming device 11 Image Reader 12 Printers 13 Paper feed section 100 control circuit 141 Main control unit 142 General Control Unit 143 OCR processing section 151 CPU 152 ROM 153 RAM 154 Image Memory 155 Image processing circuit 156 Network Communication Circuit 157 Scanner control circuit 158 Input / Output Circuit 159 Printer control circuit 160 Input / Output Circuit 161 Memory circuit 166 Bus 301 CPU 302 ROM 303 RAM 304 Memory circuit 305 GPU 306 ROM 307 RAM 308 Memory circuit 311 Network Communication Circuit 321 Main control unit 324 Recognition processing section 350 Neural Networks 350a Input layer 350b Feature Extraction Layer 350c recognition layer

Claims

1. An image correction device for correcting a photographic image formed on a sheet, comprising: image acquisition means for acquiring a photographic image; a time acquisition means for acquiring the time when the photographic image was formed, which time is estimated by artificial intelligence; a coefficient acquisition means for acquiring a correction coefficient for correcting deterioration of the photographic image according to the time elapsed since the acquired creation date; an image correction means for correcting the acquired photographic image based on the acquired correction coefficient; Equipped with The artificial intelligence uses at least information about the shape of an object in the photographic image to infer the time when the photographic image was created. An image correction device characterized by:

2. Further, the apparatus includes a recognition means having an artificial intelligence, and using the artificial intelligence, the recognition means is configured to infer the time of creation of the photographic image as a feature of an object in the acquired photographic image, The time acquisition means acquires the time when the photographic image was formed from the recognition means.

2. The image correction device according to claim 1.

3. the image correction device is connected to a server device via a network; Further, a transmitting means for transmitting the photographic image to a server device is provided, the server device has an artificial intelligence and includes a recognition means for inferring, by the artificial intelligence, the time of creation of the photographic image as a feature of an object in the received photographic image; The time acquisition means acquires the formation time from the server device.

2. The image correction device according to claim 1.

4. Furthermore, a storage means is provided for storing a correction coefficient for correcting deterioration of the photographic image in association with the elapsed time from the acquired formation time, The coefficient acquisition means acquires the correction coefficient corresponding to the elapsed time from the acquired formation time by reading it from the storage means.

4. The image correction device according to claim 2 or 3.

5. The recognition means further extracts objects and object types in the photographic image by inference; the storage means stores the correction coefficient in association with the elapsed period for each type of object, the coefficient acquisition means reads out, from the storage means, the correction coefficient corresponding to the type of the extracted object and the period of time elapsed since its creation; The image correction means corrects the extracted object based on the read correction coefficient.

5. The image correction device according to claim 4, further comprising:

6. moreover, a storage means for storing a correction coefficient for correcting deterioration of a photographic image in association with the elapsed time from the acquired formation time for each paper type of the sheet; a paper type acquisition means for acquiring the paper type of the sheet; The coefficient acquisition means acquires the correction coefficient corresponding to the elapsed time from the acquired formation time according to the acquired paper type from the storage means.

2. The image correction device according to claim 1.

7. An image forming apparatus comprising the image correction device according to claim 1, a method information acquiring unit for acquiring method information indicating an image forming method used in the image forming apparatus; The image correction device further includes a storage means for storing correction coefficients for correcting deterioration of photographic images in association with the elapsed time from the acquired formation time for each image formation method, The coefficient acquisition means acquires a correction coefficient corresponding to the elapsed time from the acquired formation time according to the acquired method information from the storage means. An image forming apparatus characterized by:

8. An image forming apparatus comprising the image correction device according to claim 1.

9. 1. An image correction method used in an image correction device that corrects a photographic image formed on a sheet, comprising: an image acquisition step of acquiring a photographic image; a time acquisition step of acquiring the time of formation of the photographic image, which is estimated by the artificial intelligence using at least information on the shape of the object in the photographic image; a coefficient acquisition step of acquiring a correction coefficient for correcting deterioration of the photographic image according to the elapsed time since the acquired formation time; an image correction step of correcting the acquired photographic image based on the acquired correction coefficient; An image correction method comprising:

Images