JP2020178168A - Image forming apparatus - Google Patents

Abstract

To provide an image forming apparatus that, when dividing an integrated image, can reduce time and effort to re-arrange divided images.SOLUTION: An image forming apparatus 100 comprises: a learning unit 218; an imaging unit 2; a division unit 211; and a first extraction unit 212. The learning unit 218 learns document data including documents to estimate relation between characters. The imaging unit 2 images a sheet formed through integration of a plurality of documents, to create imaging data. The division unit 211 divides the imaging data for each of the documents to create divided data. The divided data includes first divided data and second divided data different from the first divided data. The first extraction unit 212 extracts a first character from the first divided data and extracts a second character from the second divided data. Upon receiving input of the first character and the second character, the learning unit 218 outputs an estimation result indicating a degree of the relation between the first character and the second character.SELECTED DRAWING: Figure 2

Description

The present invention relates to an image forming apparatus.

The image processing apparatus described in Patent Document 1 includes a determination means, a first thumbnail generation means, and a display means. The determination means determines whether or not the input document is an N-up document in which the contents of the manuscript for N (N ≧ 2) pages are summarized on one page thereof. The first thumbnail generation means generates a first thumbnail which is a thumbnail of each of N pages when the determination means determines that the document is an N-up document. The display means previews the first thumbnail. The display means of the image processing apparatus described in Patent Document 1 can preview and display a 1-up document.

Japanese Unexamined Patent Publication No. 2010-28205

However, when an image in which a plurality of images are aggregated, such as an N-up document, is divided for each image, the images may not be arranged in a continuous order. Therefore, in the image forming apparatus described in Patent Document 1, the user has to take the trouble of rearranging the order of the images.

The present invention has been made in view of the above problems, and an object of the present invention is to provide an image forming apparatus capable of suppressing the trouble of rearranging the divided images when the aggregated images are divided.

The image forming apparatus according to the present invention forms a document on a sheet based on the document data indicating the document. The image forming apparatus includes a learning unit, an imaging unit, a dividing unit, and a first extraction unit. The learning unit learns document data including the document in order to estimate the connection between characters. The imaging unit captures a sheet formed by aggregating a plurality of documents to generate imaging data. The division unit divides the imaging data for each page of the document to generate a plurality of division data. The first extraction unit extracts characters from each of the divided data. The divided data includes a first divided data and a second divided data different from the first divided data. The first extraction unit extracts the first character from the first document included in the first divided data, and extracts the second character from the second document included in the second divided data. The first document indicates a document represented by the divided data for one page. The second document indicates a document represented by the divided data for one page different from the first document. The first character is a character included in either the beginning of a sentence indicating the position where the description of the first document starts or the end of the sentence indicating the position where the description of the first document ends. The second character is a character at a position different from the position including the first character at the beginning of the sentence indicating the position where the description of the second document starts or the end of the sentence indicating the position where the description of the second document ends. By inputting the first character and the second character, the learning unit outputs an estimation result indicating the degree of connection between the first character and the second character.

According to the image forming apparatus of the present invention, when the aggregated images are divided, it is possible to reduce the trouble of rearranging the divided images.

It is a figure which shows the structure of the image forming apparatus which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the control part which concerns on this Embodiment 1. It is a figure which shows the selection screen displayed on the touch panel part in Embodiment 1. It is a figure which shows the display screen which displayed the divided image which shows the divided data in Embodiment 1. It is another figure which shows the display screen which displayed the divided image which shows the divided data in Embodiment 1. FIG. It is a flowchart which shows the process which the control part executes in 1st Embodiment. It is a flowchart which shows the 1st decision process in this Embodiment 1. It is a figure which shows the structure of the control part which concerns on Embodiment 2 of this invention. It is a figure which shows the display screen which displayed the divided image which shows the divided data in Embodiment 2. It is another figure which shows the display screen which displayed the divided image which shows the divided data in Embodiment 2. It is a flowchart which shows the process which the control part executes in 2nd Embodiment. It is a flowchart which shows the 2nd decision process which a control part executes in 2nd Embodiment. It is a flowchart which shows the selection process which the control part executes in 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings, the same or corresponding parts are designated by the same reference numerals and the description is not repeated. Further, in the embodiment of the present invention, the X-axis, the Y-axis, and the Z-axis are orthogonal to each other, the X-axis and the Y-axis are parallel in the horizontal direction, and the Z-axis is parallel in the vertical direction.

[Embodiment 1]
First, the configuration of the image forming apparatus 100 according to the first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a diagram showing a configuration of an image forming apparatus 100. The image forming apparatus 100 is a color multifunction device. Further, the image forming apparatus 100 may be connected to the personal computer in a communicable manner.

As shown in FIG. 1, the image forming apparatus 100 includes an image forming unit 1, an image reading unit 2, a document conveying unit 3, and an operation display unit 4. The image forming unit 1 forms an image on the sheet P.

The image reading unit 2 reads the image formed on the sheet R and generates imaging data representing the image. Specifically, the image reading unit 2 images the sheet R and generates image data representing the sheet R. Sheet R is a manuscript. Sheet R indicates the sheet P on which the image was formed. The image contains a document. In addition, the image includes an image in which a plurality of documents are aggregated. The image reading unit 2 corresponds to an example of an “imaging unit”. The image reading unit 2 includes a contact glass, an LED (Light Emitting Mode), a mirror, a carriage, an imaging lens, and a CCD (Charge Coupled Device). The LEDs and mirrors are supported by carriages.

There are two types of image reading methods for the sheet R by the image reading unit 2: a flatbed reading mode and an ADF (Auto Document Feeder) reading mode. In the flatbed scanning mode, the image of the sheet R placed on the document placing glass is read. In the ADF reading mode, the document transport unit 3 transports the sheet R, and when the sheet R passes through the scanning position, the image of the sheet R is read. While the document transport unit 3 moves the sheet R, the CCD reads the image on the sheet R and generates imaging data representing the image. The sheet R is, for example, plain paper, copy paper, recycled paper, thin paper, thick paper, glossy paper, or an OHP (Overhead Projector) sheet.

The document transport unit 3 transports the sheet R to the image reading unit 2.

The operation display unit 4 has a touch panel 41 and operation buttons 42. The touch panel 41 has a display device 43 and a touch sensor 45. The display device 43 displays various images. The display device 43 is, for example, a liquid crystal display device (Liquid Crystal Display: LCD). The touch sensor 45 receives an operation from the user. Further, the operation button 42 accepts an operation from the user.

Each time the touch panel 41 receives an operation from the user, the touch panel 41 generates operation information indicating the content of the operation from the user. Specifically, a plurality of icons are displayed on the touch panel 41. Then, the user taps the icon. The touch panel 41 generates information indicating that the icon has been tapped as operation information.

The operation display unit 4 changes the image displayed on the touch panel 41 based on the operation from the user. Specifically, the operation display unit 4 displays on the touch panel 41 an image corresponding to the icon operated by the user among the plurality of icons displayed on the touch panel 41. Further, the operation display unit 4 reads out the information corresponding to the image displayed on the touch panel 41 from the storage unit 22.

The image forming unit 1 includes a transport mechanism 11, a feeding unit 12, a toner supply unit 13, an image forming unit 14, a fixing unit 15, a discharging unit 16, a control unit 21, and a storage unit 22. The image forming unit 1 has a transport path L.

The transport path L guides the sheet P from the feeding section 12 to the discharging section 16. The transport path L extends from the feeding section 12 to the discharging section 16.

The transport mechanism 11 transports the sheet P. Specifically, the transport mechanism 11 transports the sheet P to the discharge unit 16 via the image forming unit 14 and the fixing unit 15. Further, the transport mechanism 11 can invert the sheet P on which the image is fixed by the fixing portion 15 and convey it to the image forming portion 14.

The feeding unit 12 supplies the sheet P to the transport path L. The sheet P is, for example, plain paper, copy paper, recycled paper, thin paper, thick paper, glossy paper, or an OHP (Overhead Projector) sheet.

A plurality of toner containers are mounted on the toner supply unit 13. One of the plurality of toner containers stores cyan-colored toner. One of the plurality of toner containers stores magenta toner. One of the plurality of toner containers stores yellow toner. One of the plurality of toner containers stores black toner.

The image forming unit 14 forms an image on the sheet P. Specifically, the image forming unit 14 forms a plurality of images on the plurality of sheets P. The image forming unit 14 includes a transfer unit. The transfer unit transfers the image to the sheet P. As a result, an image is formed on the sheet P.

The image forming unit 14 includes a plurality of image forming units. One of the plurality of image forming portions forms a cyan toner image. One of the plurality of image forming portions forms a magenta toner image. One of the plurality of image forming portions forms a yellow toner image. One of the plurality of image forming portions forms a black toner image.

The image forming unit 14 includes a transfer unit, an image carrier, a charging unit, an exposure unit, and a developing unit.

The transfer unit transfers the toner image to the sheet P. The transfer section includes an intermediate transfer belt. The intermediate transfer belt is an endless belt. Toner images of a plurality of colors are formed on the intermediate transfer belt. Specifically, on the intermediate transfer belt, a plurality of image forming portions 14 form a toner image on the intermediate transfer belt. As a result, toner images of a plurality of colors are superimposed on the intermediate transfer belt, and an image is formed on the intermediate transfer belt. Then, the image formed on the intermediate transfer belt is transferred to the sheet P. As a result, an image is formed on the sheet P.

The image carrier is drum-shaped and has a rotation axis. The image carrier rotates clockwise about the axis of rotation. The image carrier has a photosensitive layer on the outer peripheral surface side.

The charged portion charges the photosensitive layer of the image carrier to a predetermined potential. The exposed portion irradiates the photosensitive layer of the image carrier with laser light to expose it. The exposed unit exposes the image carrier based on the image data. As a result, an electrostatic latent image is formed on the image carrier.

The developing unit develops an electrostatic latent image on the image carrier. The developing unit has a developing roller. The developing roller supplies toner to the image carrier and develops an electrostatic latent image on the image carrier to form a toner image. As a result, a toner image is formed on the outer peripheral surface of the image carrier.

The transfer unit transfers the toner image formed on the outer peripheral surface of the image carrier to the sheet P. As a result, the toner image is transferred to the sheet P.

The fixing portion 15 heats and pressurizes the sheet P, and fixes the image formed on the sheet P to the sheet P. Specifically, the fixing unit 15 heats and pressurizes the sheet P, and fixes the toner image formed on the sheet P on the sheet P.

The discharge unit 16 discharges the sheet P to the outside of the image forming apparatus 100. After the fixing section 15 fixes the toner image on the sheet P, the transport mechanism 11 transports the sheet P from the fixing section 15 to the discharging section 16. Then, the discharge unit 16 discharges the sheet P on which the toner image is fixed to the outside of the image forming apparatus 100.

The control unit 21 controls the operation of the image forming apparatus 100. The control unit 21 includes a processor and a storage device. The processor includes, for example, a CPU (Central Processing Unit). The processor executes a control program stored in the storage device to control the operation display unit 4, the transport mechanism 11, the feeding unit 12, the image forming unit 14, the fixing unit 15, and the discharging unit 16.

The storage unit 22 includes a storage device. Specifically, the storage unit 22 may include a memory such as a semiconductor memory, and may include an HDD (Hard Disk Drive). The storage unit 22 stores the control program.

Next, the configuration of the control unit 21 will be described in detail with reference to FIG. FIG. 2 is a diagram showing a configuration of a control unit 21 according to the present embodiment. The control unit 21 includes a learning unit 218, a division unit 211, and a first extraction unit 212. By executing the control program, the control unit 21 functions as a learning unit 218, a division unit 211, and a first extraction unit 212.

The division unit 211 divides the imaging data generated by the image reading unit 2 for each page of the document to generate a plurality of division data. The divided data includes the first divided data and the second divided data. The second divided data indicates divided data different from the first divided data.

The first extraction unit 212 extracts characters from each of the divided data. Specifically, the first extraction unit 212 extracts the first character from the first document included in the first partition data. The first document shows a document represented by one page of divided data. The first character is a character included in either the beginning or the end of the first document. Specifically, the first character is a character included in either the beginning of the sentence indicating the position where the description of the first document starts or the end of the sentence indicating the position where the description of the first document ends.

In addition, the first extraction unit 212 extracts the second character from the second document included in the second divided data. The second document shows a document represented by one page of divided data different from the first document. The second character is a character at a position different from the position including the first character at the beginning of the sentence indicating the position where the description of the second document starts or the end of the sentence indicating the position where the description of the second document ends.

Also, each of the first and second letters includes a single letter, word, and morpheme. A morpheme represents the expression element of the smallest unit that has meaning. Therefore, it can support a plurality of languages. As a result, it is possible to estimate the connection between characters in a plurality of languages.

The learning unit 218 learns the document data in order to estimate the connection between the characters. As a result, the connection between characters can be easily estimated based on the document data.

Learning includes machine learning. Machine learning includes, for example, supervised learning, unsupervised learning, and reinforcement learning. Machine learning is performed, for example, by a neural network (Neural Network) or a support vector machine (Support Vector Machine). The neural network has an input layer, a hidden layer (intermediate layer), and an output layer. The neural network uses an error backpropagation method to reduce the error between the output value and the optimum solution in the output layer.

Further, the machine learning may be deep learning. Deep learning is composed of a neural network having an input layer, two or more hidden layers, and an output layer. Specifically, deep learning is composed of, for example, a convolutional neural network, a recurrent neural network, and a Boltzmann machine.

Further, the learning unit 218 of the present embodiment outputs an estimation result indicating the degree of connection between the first character and the second character by inputting the first character and the second character. Therefore, the user does not have to reorder the document based on the characters in the document. As a result, it is possible to reduce the trouble of rearranging the documents when dividing the aggregated images.

For example, in general, when storing an aggregated document such as "2in1" used in a conference, the user uses the division function of the image forming apparatus. When a document in which a plurality of documents are aggregated is divided, the documents may not be arranged in order due to the direction in which the image reading unit 2 reads the sheet R or the like. In such cases, the user visually checks the documents, rearranges the order of the documents based on the context, and stores the documents. However, since the learning unit 218 estimates the degree of connection between characters, the user can sort the documents based on the estimation result. As a result, it is possible to reduce the trouble of rearranging the documents.

Further, the document data learned by the learning unit 218 of the present embodiment is data indicating a document formed on the sheet P by the image forming apparatus 100. The learning unit 218 learns the document data of the documents that the image forming apparatus 100 aggregates a plurality of documents and forms on one sheet P. Therefore, since the first character and the second character input to the learning unit 218 are included in the document data, the learning accuracy of the learning unit 218 is improved. As a result, the accuracy of the estimation result is improved.

For example, the image forming apparatus 100 forms a document to be used in a conference on a sheet P. Specifically, the image forming apparatus 100 forms a document to be used in the conference on the sheet P based on the document data input to the image forming apparatus 100. A plurality of documents are aggregated on the sheet P. Then, the learning unit 218 learns the document data. After that, the image reading unit 2 reads the sheet R and generates imaging data. Sheet R is a sheet P in which a plurality of documents are aggregated. Further, the divided data is generated based on the document of the imaging data. Further, the first character and the second character acquired from the divided data are input to the learning unit 218. That is, the first character and the second character input to the learning unit 218 are included in the learning data already learned by the learning unit 218. Therefore, the learning unit 218 can output an accurate estimation result. The learning data is stored in the storage unit 22.

The document data learned by the learning unit 218 may include the document data input to the image forming apparatus 100. Therefore, the learning unit 218 can learn the image data of the document that is not actually formed on the sheet P.

In addition, the learning unit 218 of the present embodiment further learns the document data and the page number corresponding to the document data. As a result, the learning accuracy is improved, and the accuracy of estimating the connection between characters is improved.

In addition, the learning unit 218 executes natural language processing on the document data. Natural language processing is a series of technologies that allow a computer to process natural language. Natural language is the language used by humans for communication. Natural language processing includes morphological analysis, parsing, semantic analysis and context analysis.

The learning unit 218 executes morphological analysis on the document data. Morphological analysis is a process of classifying morphemes into the smallest unit of meaning. The learning unit 218 classifies the text included in the document data into morphemes based on the dictionary data stored in the storage unit 22. Then, the learning unit 218 executes the syntactic analysis based on the result of the morphological analysis. Parsing is the process of analyzing the relationship between morphemes. The relationship is, for example, a modification relationship between a morpheme and a morpheme. Further, the learning unit 218 executes the semantic analysis based on the result of the syntactic analysis. Semantic analysis is a process of determining a syntax tree based on the result of parsing. The syntax tree shows the progress and results of parsing in a tree structure. Further, the learning unit 218 executes the context analysis based on the result of the semantic analysis. Context analysis is a process of analyzing the relationship between sentences. The sentence contains the subject and the predicate and indicates one complete statement. The learning unit 218 learns based on the result of natural language processing. The result of natural language processing is stored in the storage unit 22 as learning data.

Subsequently, the configuration of the control unit 21 will be described in detail with reference to FIG. The control unit 21 further includes a first determination unit 213. The control unit 21 functions as the first determination unit 213 by executing the control program.

The first determination unit 213 determines the order of the divided data. Specifically, the first determination unit 213 determines the order of the divided data based on the estimation result of the learning unit 218. More specifically, the first determination unit 213 inputs the first character and the second character to the learning unit 218, and estimates that the degree of connection between the first character and the second character output by the learning unit 218 is indicated. Based on the result, the order of the first divided data and the second divided data is determined.

Therefore, the user does not have to reorder the document based on the characters in the document. As a result, it is possible to reduce the trouble of rearranging the documents when dividing the aggregated images.

Further, the first extraction unit 212 of the first embodiment extracts the first character located at the end of the sentence of the first document and extracts the second character located at the beginning of the sentence of the second document. Then, the first determination unit 213 inputs the first character and the second character to the learning unit 218, and based on the estimation result output from the learning unit 218, the order of the first divided data and the second divided data. To determine. Therefore, the first document and the second document can be read in succession. As a result, the aggregated and disjointed documents can be made into one connected document.

In addition, the first extraction unit 212 of the first embodiment extracts the first character and the second character. The first character is located at the end of the first document. The second character is located at the beginning of the second document. Then, the first determination unit 213 inputs the first character and the second character to the learning unit 218, and based on the estimation result output from the learning unit 218, the order of the first divided data and the second divided data. To determine. Therefore, the first document and the second document can be read in succession. As a result, the aggregated and disjointed documents can be made into one connected document.

Next, the selection screen 50 displayed on the operation display unit 4 will be described with reference to FIGS. 2 and 3. FIG. 3 is a diagram showing a selection screen 50 displayed on the touch panel unit 41. The selection screen 50 is displayed on the touch panel unit 41 by operating the operation button 42 for selecting the copy function of the image forming apparatus 100. On the selection screen 50, the paper selection icon 51, the reduction / enlargement icon 52, the density setting icon 53, the double-sided / split setting icon 54, the page aggregation setting icon 55, the sort / sorting setting icon 56, the function list icon 57, and favorites The icon 58 is displayed.

The paper selection icon 51 is operated by the user when selecting the size of the sheet P. The reduction / enlargement icon 52 is operated by the user when the image formed on the sheet P is enlarged or reduced. The density setting icon 53 is operated by the user when setting the copy density. The double-sided / split setting icon 54 is operated by the user when the double-sided or single-sided setting and an image in which a plurality of images such as "2in1" are aggregated are divided into images. The page aggregation setting icon 55 is operated by the user when setting page aggregation such as "2in1". The sort / sorting setting icon 56 is operated by the user when setting sorting conditions such as the presence / absence of sorting. The function list icon 57 is operated by the user when displaying a function list screen explaining various functions on the touch panel 41. The favorite icon 58 is operated by the user when displaying an icon frequently used by the user on the touch panel 41.

Further, as shown in FIG. 3, the double-sided / split setting icon 54 is tapped by the finger (for example, the index finger) of the user H1's hand. In response to this operation, the touch panel unit 41 displays the screen corresponding to the double-sided / split setting icon 54. The hand is not displayed on the touch panel 41.

Next, the screen corresponding to the double-sided / split setting icon 54 will be described with reference to FIGS. 2 to 4. FIG. 4 is a diagram showing a display screen 110 displaying a divided image D showing divided data. As shown in FIG. 4, the display screen 110 includes a first display area 111 and a second display area 112.

In the first display area 111, a preview image 113 for displaying the captured image RG1 showing the captured data and a return button 114 are displayed. The captured image RG1 shown in FIG. 4 is a “2in1” image in which two images are aggregated on one sheet R. When the captured image RG1 includes a plurality of documents, the captured image RG1 is also displayed in a plurality. The back button 114 is a button for returning to the selection screen 50 shown in FIG.

In the second display area 112, a plurality of divided images D showing the divided data are displayed. Each of the plurality of divided images D indicates an image represented by one page of divided data. The plurality of divided images D shown in FIG. 4 include a first divided image D1 and a second divided image D2.

The first divided image D1 is, for example, an image showing a document on the second page of a plurality of pages. The first divided image D1 includes a document. The first divided image D1 includes a first sentence beginning region BS1 and a first sentence ending region ES1.

The first sentence beginning area BS1 indicates an area in which the sentence beginning portion is located in the document included in the first divided image D1. The beginning of the sentence includes the position where the description of the document starts in the document included in the image represented by the divided data for one page. In the first sentence head region BS1 shown in FIG. 4, the character string "multifaction devices ..." is located.

The first sentence end area ES1 indicates an area in which the sentence end portion is located in the document included in the first divided image D1. The end of the sentence includes the position where the description of the document ends in the document included in the image represented by the divided data for one page. The character string "... optimally" is located in the first sentence end region ES1 shown in FIG.

The second divided image D2 is, for example, an image showing a document on the first page of a plurality of pages. The second divided image D2 includes a document. The second divided image D2 includes the second sentence beginning region BS2 and the second sentence ending region ES2.

The second sentence beginning area BS2 indicates an area in which the sentence beginning portion is located in the document included in the second divided image D2. The character string "Tokyo Co., Ltd ....." is located in the second sentence head region BS2 shown in FIG.

The second sentence end area ES2 indicates an area in which the sentence end portion of the document included in the second divided image D2 is located. The character string "... printers and" is located in the second sentence end region ES2 of the second divided image D2 shown in FIG.

Further, the first divided image D1 shown in FIG. 4 is located on the upstream side in the direction from the first display area 111 to the second display area 112. The second divided image D2 shown in FIG. 4 is located on the downstream side in the direction from the first display area 111 to the second display area 112. Therefore, in FIG. 4, the first divided image D1 on the second page is located on the upstream side, and the second divided image D2 on the first page is located on the downstream side.

When the captured image RG1 is divided into the first divided image D1 shown in FIG. 4 and the second divided image D2 shown in FIG. 4, the control unit 21 displays the display screen 110 on which the operation display unit 4 sets the division. In addition, the operation display unit 4 is controlled. When the display screen 110 shown in FIG. 4 is displayed on the touch panel 41 of the operation display unit 4, the double-sided / split setting icon 54 shown in FIG. 3 is touched twice. When the touch operation is performed once, the double-sided setting screen is displayed on the operation display unit 4. When the division setting icon 54 is touch-operated twice, the image reading unit 2 images the sheet R and acquires the imaging data representing the sheet R. The sheet R is a sheet on which an image is formed. The image formed on the sheet R is an image in which a plurality of images are aggregated. The imaged data captured by the image reading unit 2 is transmitted to the control unit 21.

The control unit 21 receives the imaging data. Then, the control unit 21 acquires the brightness of a predetermined region of the imaging data. The predetermined region indicates a region formed between images adjacent to each other when a plurality of images are aggregated. In addition, the pattern of a predetermined area is changed according to the number of images to be aggregated. When there are two aggregated images, the pattern of the predetermined region is, for example, the shape of a straight line that divides the captured image into two. When there are four aggregated images, the pattern of the predetermined region is, for example, a cross shape that divides the captured image into four.

Further, the control unit 21 determines whether or not the brightness of the predetermined region has a predetermined gradation. The predetermined gradation indicates, for example, white. Then, the dividing unit 211 divides the imaging data based on a predetermined region showing the gradation of white. For example, the division unit 211 divides the imaging data into the first division data and the second division data.

Further, the control unit 21 executes character recognition processing on the divided data. The character recognition process is typically an optical character recognition (OCR) process. The control unit 21 executes a character recognition process on the divided data to search for a character image. Specifically, the control unit 21 executes character recognition processing on the divided data from a plurality of directions. Therefore, the accuracy of searching the character image is improved. As a result, the control unit 21 can acquire accurate text information based on the searched character image.

Further, when a plurality of images are aggregated, the orientation of the image data is changed and the images are aggregated. Then, the orientation of the divided data of the dividing unit 211 is the same as the orientation when the captured data is captured by the image reading unit 2. Therefore, the orientation of the document contained in the divided data is different from the orientation of generally recognized characters. The control unit 21 performs character recognition processing on the divided data from a plurality of directions, and corrects the direction of the divided data in the direction in which the most character images can be searched. As a result, when the user confirms the divided data, it becomes easy to recognize the divided data.

Then, the control unit 21 controls the operation display unit 4 so that the first divided image D1 showing the first divided data is displayed in the second display area 112 shown in FIG. The control unit 21 controls the operation display unit 4 so that the second divided image D2 showing the first divided data is displayed in the second display area 112 shown in FIG.

Next, with reference to FIGS. 2 to 5, the process until the first determination unit 213 determines the order of the first division data and the second division data will be described. FIG. 5 is another view showing the display screen 110 displaying the divided image D showing the divided data. As shown in FIG. 5, the display screen 110 includes a first display area 111 and a second display area 112.

In the first display area 111, the captured image RG1 showing the captured data is displayed. The captured image RG1 includes a preview image 113 and a back button 114. In the second display area 112, a plurality of divided images D showing the divided data are displayed. The plurality of divided images D shown in FIG. 5 include a first divided image D1 and a second divided image D2.

The first divided image D1 shown in FIG. 5 is located on the downstream side in the direction from the first display area 111 to the second display area 112. The second divided image D2 shown in FIG. 5 is located on the upstream side in the direction from the first display area 111 to the second display area 112. Therefore, in FIG. 5, the second divided image D2 on the first page is located on the upstream side, and the first divided image D1 on the second page is located on the downstream side. That is, the page numbers are arranged in ascending order from the upstream side.

Further, the characters located in the second sentence end region ES2 of the second divided image D2 and the characters located in the first sentence beginning region BS1 of the first divided image D1 have meanings by connecting the characters. It is a character string to be formed. Specifically, as shown in FIG. 5, the word "and" located in the second sentence end region ES2 and the word "multifaction" located in the first sentence beginning region BS1 of the first divided image D1 are "multifunction". It is a sequence of words "and partition".

As shown in FIG. 5, when connecting characters to form a meaningful character string, the first extraction unit 212 extracts characters from each of the divided data. For example, the first extraction unit 212 extracts a character or a character string based on the text data generated by the control unit 21 for each of the divided data. Specifically, the first extraction unit 212 extracts the word "multibunction" from the first sentence head region BS1 of the document included in the first divided image D1 shown in FIG. The document included in the first divided image D1 corresponds to an example of the "first document". The word extracted from the first sentence head area BS1 corresponds to an example of the "first character". Then, the first extraction unit 212 extracts the word "and" from the second sentence end region ES2 of the document included in the second divided image D2 shown in FIG. The document included in the second divided image D2 corresponds to an example of the "second document". The word extracted from the second sentence end region ES2 corresponds to an example of the "second character".

Then, the first determination unit 213 inputs the first character and the second character to the learning unit 218. Further, the learning unit 218 outputs an estimation result indicating the degree of connection between the first character and the second character. For example, “and” of the second sentence end region ES2 and “multifunction” of the first sentence beginning region BS1 are input to the learning unit 218. Then, the learning unit 218 outputs an estimation result indicating the degree of connection between the “and” and the “multifunction”.

Further, the first extraction unit 212 may extract characters from the first sentence end region ES1 of the first divided image D1 and the second sentence beginning region BS2 of the second divided image D2 shown in FIG. Specifically, the first extraction unit 212 extracts the word "optimally" from the first sentence end region ES1 of the first divided image D1 shown in FIG. The first extraction unit 212 extracts the word "Tokyo" from the second sentence head region BS2 of the second divided image D2 showing the second divided data shown in FIG.

Then, the first determination unit 213 inputs the first character and the second character to the learning unit 218. Further, the learning unit 218 outputs an estimation result indicating the degree of connection between the first character and the second character. For example, the learning unit 218 is input with "optimally" of the first sentence end region ES1 and "Tokkyo" of the second sentence beginning region BS2. Then, the learning unit 218 outputs an estimation result indicating the degree of connection between the “optimally” and the “Tokyo”.

Then, the first determination unit 213 determines the order of the first division data and the second division data based on the estimation result indicating the degree of connection between the first character and the second character output by the learning unit 218. .. Specifically, the estimation result indicating the degree of connection between "and" and "multifaction" is compared with the estimation result indicating the degree of connection between "optimally" and "Tokkyo", and the degree of connection is estimated to be large. Based on the result, the first determination unit 213 determines the order of the first division data and the second division data.

Further, the control unit 21 controls the operation display unit 4 so that the operation display unit 4 displays the first divided image D1 and the second divided image D2 based on the determination of the first determination unit 213. Therefore, as shown in FIG. 5, the first divided image D1 and the second divided image D2 are arranged in the order of pages. As a result, the first divided image D1 and the second divided image D2 can be read continuously.

Further, the estimation result of the learning unit 218 and the determination result of the first determination unit 213 are learned by the learning unit 218. Therefore, the learning unit 218 relearns with the document data, the page number, the estimation result, and the determination result. As a result, an accurate estimation result can be output.

Next, the process executed by the control unit 21 of the first embodiment will be described with reference to FIG. FIG. 6 shows a flowchart of processing executed by the control unit 21. The process executed by the control unit 21 includes steps S101 to S108.

In step S101, the control unit 21 controls the operation display unit 4 so that the operation display unit 4 displays the selection screen 50. The process proceeds to step S102.

In step S102, the control unit 21 determines whether or not the signal acquired from the operation display unit 4 includes an instruction to divide the image data. If the instruction to divide the image data is not included (No in step S102), the process ends. When the instruction to divide the image data is included (Yes in step S102), the process proceeds to step S103.

In the case of Yes in step S102, in step S103, the control unit 21 acquires the image pickup data generated by the image reading unit 2. The process proceeds to step S104.

In step S104, the division unit 211 divides the imaging data for each document to generate the divided data. The process proceeds to step S105.

In step S105, the control unit 21 executes a character image search for the divided data and acquires text information corresponding to the document. The process proceeds to step S106.

In step S106, the control unit 21 corrects the orientation of the divided data based on the acquisition rate of the character image. Specifically, the control unit 21 corrects the orientation of the divided data in the orientation in which the most character images can be searched. The process proceeds to step S107.

In step S107, the control unit 21 executes the first determination process. The first determination process will be described later with reference to FIG. 7. The process proceeds to step S108.

In step S108, the learning unit 218 learns the document data, the page number, the estimation result, and the determination result. The process ends.

Next, the first determination process executed by the control unit 21 will be described with reference to FIG. 7. FIG. 7 is a diagram showing a flowchart of the first determination process. The first determination process includes steps S201 to S210. The first determination process shown in FIG. 7 corresponds to step S107 shown in FIG.

In step S201, the first extraction unit 212 extracts the first character from the first sentence end region ES1 of the first document included in the first partition data. The process proceeds to step S202.

In step S202, the first extraction unit 212 extracts the second character from the second sentence beginning region BS2 of the second document included in the second divided data. The process proceeds to step S203.

In step S203, the first determination unit 213 inputs the first character located at the end of the sentence of the first document and the second character located at the beginning of the sentence of the second document to the learning unit 218. The process proceeds to step S204.

In step S204, the learning unit 218 outputs an estimation result indicating the degree of connection between the first character and the second character. The process proceeds to step S205.

In step S205, the first extraction unit 212 extracts the first character from the first sentence head region BS1 of the first document included in the first partition data. The process proceeds to step S206.

In step S206, the first extraction unit 212 extracts the second character from the second sentence end region ES2 of the second document included in the second divided data. The process proceeds to step S207.

In step S207, the first determination unit 213 inputs the first character located at the beginning of the sentence of the first document and the second character located at the end of the sentence of the second document to the learning unit 218. The process proceeds to step S208.

In step S208, the learning unit 218 outputs an estimation result indicating the degree of connection between the first character and the second character. The process proceeds to step S209.

In step S209, the control unit 21 determines whether or not there is other divided data. If there is other divided data (Yes in step S209), the process returns to step S201. If there is no other divided data (No in step S209), the process proceeds to step S210.

If No in step S209, in step S210, the first determination unit 213 determines the order of the first division data and the second division data. The process returns to step S108 shown in FIG.

[Embodiment 2]
Next, the image forming apparatus 100 of the second embodiment will be described with reference to FIG. The image forming apparatus 100 of the second embodiment is different from the image forming apparatus 100 of the first embodiment in that it has a second extraction unit 214, a second determination unit 215, a determination unit 216, and a selection unit 217. Hereinafter, the items different from the first embodiment will be described with respect to the second embodiment, and the description of the parts overlapping with the first embodiment will be omitted.

FIG. 8 is a diagram showing the configuration of the control unit 21 of the second embodiment. The control unit 21 includes a division unit 211, a first extraction unit 212, a first determination unit 213, a learning unit 218, a second extraction unit 214, and a second determination unit 215. By executing the control program, the control unit 21 functions as a division unit 211, a first extraction unit 212, a first determination unit 213, a learning unit 218, a second extraction unit 214, and a second determination unit 215. The division unit 211, the first extraction unit 212, the first determination unit 213, and the learning unit 218 are the same as those in the first embodiment, and thus the description thereof will be omitted.

The second extraction unit 214 extracts a symbol located in a predetermined area of the document. Symbols include letters and numbers. A predetermined area of the document includes a header area of the document or a footer area of the document. Therefore, the second extraction unit 214 can acquire the page number attached to the document.

The second determination unit 215 determines the order of the first division data and the second division data based on the extraction result of the second extraction unit 214. The result of the second determination unit 215 can be acquired in accordance with the result of the first determination unit 213. As a result, the order of the first divided data and the second divided data can be accurately determined.

Next, a process in which the second determination unit 215 determines the order of the first division data and the second division data will be described with reference to FIGS. 8 to 10. FIG. 9 is a diagram showing a display screen 110 displaying a divided image D showing divided data. As shown in FIG. 9, the display screen 110 includes a first display area 111 and a second display area 112.

In the first display area 111, a preview image 113 for displaying the captured image RG2 showing the captured data and a return button 114 are displayed. The captured image RG2 shown in FIG. 9 is a “4in1” image in which four images are aggregated on one sheet R.

In the second display area 112, a plurality of divided images D showing the divided data are displayed. The plurality of divided images D shown in FIG. 9 include a first divided image D1, a second divided image D2, a third divided image D3, and a fourth divided image D4.

The first divided image D1 is, for example, an image showing a document on the third page of a plurality of pages. The first divided image D1 includes a document. The first divided image D1 includes a first sentence beginning region BS1, a first sentence ending region ES1, and a first extraction region CT1. The character string "arranges ..." is located in the first sentence head region BS1 of the first divided image D1 shown in FIG. The character string "... has products" is located in the first sentence end region ES1 of the first divided image D1 shown in FIG. The symbol "3" is located in the first extraction region CT1 of the first divided image D1 shown in FIG.

The second divided image D2 is, for example, an image showing a document on the fourth page of a plurality of pages. The second divided image D2 includes a document. The second divided image D2 includes the second sentence beginning region BS2, the second sentence ending region ES2, and the second extraction region CT2. The character string "comprehensive services ..." is located in the second sentence head region BS2 of the second divided image D2 shown in FIG. The character string "... in the UK." Is located in the second sentence end region ES2 of the second divided image D2 shown in FIG. The symbol "4" is located in the second extraction region CT2 of the second divided image D2 shown in FIG.

The third divided image D3 is, for example, an image showing a document on the first page of a plurality of pages. The third divided image D3 includes a document. The third divided image D3 includes a third sentence beginning region BS3, a third sentence ending region ES3, and a third extraction region CT3. The character string "Tokyo Co., Ltd ...." is located in the third sentence head region BS3 of the third divided image D3 shown in FIG. The character string "... printers and" is located in the third sentence end region ES3 of the third divided image D3 shown in FIG. The symbol "1" is located in the third extraction region CT3 of the third divided image D3 shown in FIG.

The fourth divided image D4 is, for example, an image showing a document on the second page of a plurality of pages. The fourth divided image D4 includes a document. The fourth divided image D4 includes the fourth sentence beginning region BS4, the fourth sentence ending region ES4, and the fourth extraction region CT4. The character string "multifunction devices ..." is located in the fourth sentence head region BS4 of the fourth divided image D4 shown in FIG. The character string "... optimally" is located in the fourth sentence end region ES4 of the fourth divided image D4 shown in FIG. The symbol "2" is located in the fourth extraction region CT4 of the fourth divided image D4 shown in FIG.

Further, the first divided image D1 and the third divided image D3 shown in FIG. 9 are located on the upstream side in the direction from the first display area 111 to the second display area 112. The second divided image D2 and the fourth divided image D4 are located on the downstream side in the direction from the first display area 111 to the second display area 112. Therefore, the first divided image D1 showing the third page is located upstream of the fourth divided image D4 showing the second page.

Further, the first divided image D1 and the second divided image D2 shown in FIG. 9 are located on the upstream side in the direction from the preview image 113 toward the return button 114. The third divided image D3 and the fourth divided image D4 are located on the downstream side in the direction from the preview image 113 toward the return button 114. Therefore, the first divided image D1 showing the third page and the second divided image D2 showing the fourth page are upstream of the third divided image D3 showing the first page and the fourth divided image D4 showing the second page. To position. Therefore, the first divided image D1 showing the third page is located upstream of the third divided image D3 showing the first page. That is, the first divided images D1 to the fourth divided images D4 shown in FIG. 9 are not arranged in order.

FIG. 10 is another diagram showing a display screen 110 displaying a divided image D showing divided data. As shown in FIG. 10, the display screen 110 includes a first display area 111 and a second display area 112.

In the first display area 111, a preview image 113 for displaying the captured image RG2 showing the captured data and a return button 114 are displayed. In the second display area 112, a plurality of divided images D showing the divided data are displayed. The plurality of divided images D shown in FIG. 10 include a first divided image D1, a second divided image D2, a third divided image D3, and a fourth divided image D4.

Further, the third divided image D3 and the first divided image D1 shown in FIG. 10 are located on the upstream side in the direction from the first display area 111 to the second display area 112. The fourth divided image D4 and the second divided image D2 are located on the downstream side in the direction from the first display area 111 to the second display area 112. Therefore, the third divided image D3 showing the first page is located upstream of the fourth divided image D4 showing the second page. Further, the first divided image D1 showing the third page is located upstream of the second divided image D2 showing the fourth page.

Further, the third divided image D3 and the fourth divided image D4 shown in FIG. 10 are located on the upstream side in the direction from the preview image 113 toward the return button 114. The first divided image D1 and the second divided image D2 shown in FIG. 10 are located on the downstream side in the direction from the preview image 113 toward the return button 114. Therefore, the fourth divided image D4 showing the second page is located upstream of the first divided image D1 and the second divided image D2. That is, the first divided images D1 to the fourth divided images D4 shown in FIG. 10 are arranged in ascending order.

As shown in FIG. 10, when the first divided images D1 to the fourth divided images D4 are arranged in ascending order, the second extraction unit 214 extracts symbols from each of the divided data. For example, the second extraction unit 214 extracts characters based on the text data generated by the control unit 21 for each of the divided data. Specifically, the second extraction 214 extracts the number "3" from the first extraction region CT1 of the first divided image D1 shown in FIG. The second extraction 214 extracts the number “4” from the second extraction region CT2 of the second divided image D2. The second extraction 214 extracts the number "1" from the third extraction region CT3 of the third divided image D3. The second extraction 214 extracts the number “2” from the fourth extraction region CT4 of the fourth divided image D4.

Then, the second determination unit 215 determines the order of the first division data, the second division data, the third division data, and the fourth division data based on the extraction result of the second extraction unit 214. Therefore, the 1-division data, the 2nd division data, the 3rd division data, and the 4th division data are arranged in order. As a result, the user can reduce the trouble of arranging the divided data.

Subsequently, with reference to FIGS. 9 and 10, the processing of the first extraction unit 212 and the first determination unit 213 when the divided data is two or more will be described. The control unit 21 of the second embodiment can acquire the determination result of the first determination unit 213 and the determination result of the second determination unit 215.

The first extraction unit 212 extracts characters based on the text data generated by the control unit 21 for each of the divided data. Specifically, the first extraction unit 212 extracts the word "arranges" from the first sentence head region BS1 of the document included in the first divided image D1 shown in FIG. The document included in the first divided image D1 corresponds to an example of the "first document". The word extracted from the first sentence head area BS1 corresponds to an example of the "first character".

Then, the first extraction unit 212 extracts the word "UK." From the second sentence end region ES2 of the document included in the second divided image D2 shown in FIG. The document included in the second divided image D2 corresponds to an example of the "second document". The word extracted from the second sentence end region ES2 corresponds to an example of the "second character".

Then, the first extraction unit 212 extracts the word "and" from the third sentence end region ES3 of the document included in the third divided image D3 shown in FIG. The document included in the third divided image D3 corresponds to an example of the "second document". The word extracted from the third sentence end region ES3 corresponds to an example of the "second character".

Then, the first extraction unit 212 extracts the word "optimally" from the fourth sentence end region ES4 of the document included in the fourth divided image D4 shown in FIG. The document included in the fourth divided image D4 corresponds to an example of the "second document". The word extracted from the fourth sentence end region ES4 corresponds to an example of the "second character".

Then, the first determination unit 213 inputs the first character and the second character to the learning unit 218. Further, the learning unit 218 outputs an estimation result indicating the degree of connection between the first character and the second character.

For example, "UK." In the second sentence end region ES2 and "arranges" in the first sentence beginning region BS1 are input to the learning unit 218. Then, the learning unit 218 outputs the first estimation result indicating the degree of connection between "UK." And "arranges".

For example, “and” of the third sentence end region ES3 and “arranges” of the first sentence beginning region BS1 are input to the learning unit 218. Then, the learning unit 218 outputs a second estimation result indicating the degree of connection between "and" and "arranges".

For example, “optimally” of the fourth sentence end region ES4 and “arranges” of the first sentence beginning region BS1 are input to the learning unit 218. Then, the learning unit 218 outputs a third estimation result indicating the degree of connection between the “optimally” and the “arranges”.

Then, the first determination unit 213 determines the order of the first division data and the second division data based on the estimation result indicating the degree of connection between the first character and the second character output by the learning unit 218. .. Specifically, the first estimation result, the second estimation result, and the third estimation result are compared, and the first determination unit 213 performs the first division data and the second division based on the estimation result having a large degree of connection. Determine the order with the data.

Further, the first extraction unit 212 repeats the same processing for each divided image D. Further, each time the first extraction unit 212 executes the process, the first determination unit 213 determines the order of the first division data and the second division data. Then, the control unit 21 displays the first divided image D1 to the fourth divided image D4 in the order of pages, as shown in FIG. 10, based on the determination result of the first determination unit 213. As a result, the first divided image D1 to the fourth divided image D4 can be continuously read.

Subsequently, the control unit 21 of the second embodiment will be described in more detail with reference to FIGS. 8 to 10. The control unit 21 further includes a determination unit 216 and a selection unit 217. The control unit 21 functions as a determination unit 216 and a selection unit 217 by executing a control program.

The determination unit 216 determines whether or not the determination result of the first determination unit 213 and the determination result of the second determination unit 215 match. Therefore, when the determination result of the first determination unit 213 and the determination result of the second determination unit 215 match, it can be determined that the accuracy of the first determination unit 213 and the second determination unit 215 is high. If the determination result of the first determination unit 213 and the determination result of the second determination unit 215 do not match, it can be determined that the accuracy of either the first determination unit 213 or the second determination unit 215 is low. As a result, the superiority or inferiority of the first determination unit 213 and the second determination unit 215 can be determined by using the determination result of the determination unit 216 as a trigger.

The selection unit 217 selects the determination result of either the determination result of the first determination unit 213 or the determination result of the second determination unit 215. Specifically, when the determination unit 216 determines that the determination results match, the selection unit 217 selects the determination result of the first determination unit 213. When the determination unit 216 determines that the determination results do not match, the selection unit 217 selects one of the determination result of the first determination unit 213 and the determination result of the second determination unit 215. .. When the first result and the second result are different, one of the first result and the second result has priority. Therefore, the result of the determination unit that can accurately determine the order of the first document and the second document can be adopted. As a result, the user can reduce the trouble of determining the order of the first divided data and the second divided data.

For example, when the determination results do not match, the user stores in advance in the storage unit 22 that the determination result of the second determination unit 215 is selected. Therefore, when the determination unit 216 determines that the determination results do not match, the selection unit 217 selects the determination result of the second determination unit 215. As a result, when the learning of the learning unit 218 has not progressed, the order of the divided data can be accurately determined by selecting the determination result of the second determination unit 215.

Further, when the learning of the learning unit 218 progresses, the storage unit 22 stores in advance that the user selects the determination result of the first determination unit 213. Therefore, when the determination unit 216 determines that the determination results do not match, the selection unit 217 selects the determination result of the first determination unit 213. As a result, since the determination result of the first determination unit 213 determined based on the estimation result with advanced learning and improved accuracy can be selected, the order of the divided data can be determined with high accuracy.

Next, the process executed by the control unit 21 of the second embodiment will be described with reference to FIG. FIG. 11 shows a flowchart of processing executed by the control unit 21. The process executed by the control unit 21 includes steps S301 to S310. Steps S301 to S307 shown in FIG. 11 correspond to steps S101 to S107 shown in FIG. 6, and the same processing is executed.

After step S307, in step S308, the control unit 21 executes the second determination process. The second determination process will be described later with reference to FIG. The process proceeds to step S309.

In step S309, the control unit 21 executes the selection process. The selection process will be described later with reference to FIG. The process proceeds to step S310.

In step S310, the learning unit 218 learns the document data, the page number, the estimation result, the determination result of the first determination unit 213, and the determination result of the second determination unit 215. The process ends.

Next, the second determination process will be described with reference to FIG. FIG. 12 shows a flowchart of the second determination process executed by the control unit 21. The second determination process executed by the control unit 21 includes steps S401 to S403. The second determination process corresponds to step S308 shown in FIG.

In step S401, the second extraction unit 214 extracts the symbol located in the extraction area CT of the document included in the divided data. The process proceeds to step S402.

In step S402, the control unit 21 determines whether or not there is other divided data. If there is other divided data (Yes in step S402), the process returns to step S401. If there is no other divided data (No in step S402), the process proceeds to step S403.

In step S403, the second determination unit 215 determines the order of the first division data and the second division data. The process returns to step S309 shown in FIG.

Next, the selection process will be described with reference to FIG. FIG. 13 shows a flowchart of the selection process executed by the control unit 21. The selection process executed by the control unit 21 includes steps S501 to S503. The selection process corresponds to step S309 shown in FIG.

In step S501, the determination unit 216 determines whether or not the determination result of the first determination unit 213 and the determination result of the second determination unit 215 match. If the determination results match (Yes in step S501), the process proceeds to step S502. If the determination results do not match (No in step S501), the process proceeds to step S503.

If No in step S502, in step S503, the selection unit 217 selects a predetermined determination result. The process returns to step S310 shown in FIG.

In the case of Yes in step S502, in step S502, the selection unit 217 selects the determination result of the first determination unit 213. The process returns to step S310 shown in FIG.

The embodiments of the present invention have been described above with reference to the drawings. However, the present invention is not limited to the above-described embodiment, and can be implemented in various embodiments without departing from the gist thereof. In addition, various inventions can be formed by appropriately combining the plurality of components disclosed in each of the above embodiments. For example, some components may be removed from all the components shown in the embodiments. Further, components over different embodiments may be combined as appropriate. In order to make the drawings easier to understand, each component is schematically shown, and the thickness, length, number, spacing, etc. of each component shown are actual for the convenience of drawing creation. Is different. Further, the speed, material, shape, dimensions, etc. of each component shown in the above embodiment are merely examples, and are not particularly limited, and various changes can be made without substantially deviating from the configuration of the present invention. It is possible.

(1) When the determination unit 216 determines that the determination results do not match, the selection unit 217 of the second embodiment selects the determination result of the predetermined determination unit. However, if the determination results do not match, the control unit 21 may let the user determine the order of the divided data. Then, the learning unit 218 learns the order determined by the user. Therefore, the learning unit 218 can learn with high accuracy. As a result, the learning unit 218 can output an accurate estimation result.

The present invention can be used in the field of image forming apparatus.

2 Image reading unit (imaging unit)
21 Control unit 100 Image forming device 211 Dividing unit 212 First extraction unit 213 First determination unit 214 Second extraction unit 214 Second extraction 215 Second determination unit 216 Judgment unit 217 Selection unit 218 Learning unit CT extraction area (predetermined area)
P sheet R sheet

Claims (10)

An image forming apparatus that forms a document on a sheet based on document data indicating a document.
A learning unit that learns document data including the above documents in order to estimate the connection between characters,
An imaging unit that generates imaging data by imaging a sheet formed by aggregating multiple documents,
A division unit that divides the imaging data into one page of the document to generate a plurality of division data, and
A first extraction unit that extracts characters from each of the divided data,
With
The divided data includes a first divided data and a second divided data different from the first divided data.
The first extraction unit
The first character is extracted from the first document included in the first partition data,
The second character is extracted from the second document included in the second divided data, and
The first document indicates a document represented by the divided data for one page.
The second document indicates a document represented by the divided data for one page different from the first document.
The first character is a character included in either the beginning of a sentence indicating the position where the description of the first document starts or the end of the sentence indicating the position where the description of the first document ends.
The second character is a character at a position different from the position including the first character in the beginning of the sentence indicating the position where the description of the second document starts or the end of the sentence indicating the position where the description of the second document ends.
The learning unit is an image forming apparatus that outputs an estimation result indicating the degree of connection between the first character and the second character by inputting the first character and the second character. The image forming apparatus according to claim 1, wherein the document data learned by the learning unit is data including the document formed by the image forming apparatus on a sheet. The image forming apparatus according to claim 1 or 2, wherein the learning unit further learns the document data and the page number corresponding to the document data. The image forming apparatus according to claim 1 or 2, wherein each of the first character and the second character contains a single character, a word, and a morpheme. The first extraction unit
The first character located at the end of the first document is extracted and
The image forming apparatus according to any one of claims 1 to 4, which extracts the second character located at the beginning of the sentence of the second document. The first extraction unit
The first character located at the beginning of the first document is extracted and
The image forming apparatus according to any one of claims 1 to 5, which extracts the second character located at the end of the sentence of the second document. The invention according to any one of claims 1 to 6, further comprising a first determination unit that determines the order of the first division data and the second division data based on the estimation result of the learning unit. Image forming device. A second extraction unit that extracts symbols located in a predetermined area of the document, and
The image forming apparatus according to claim 7, further comprising a second determination unit that determines the order of the first division data and the second division data based on the extraction result of the second extraction unit. A determination unit that determines whether or not the determination result of the first determination unit and the determination result of the second determination unit match.
Further provided with a selection unit for selecting the determination result of the first determination unit or the determination result of the second determination unit.
When the determination unit determines that the determination result of the first determination unit and the determination result of the second determination unit do not match, the selection unit determines the determination result of the first determination unit and the determination of the second determination unit. The image forming apparatus according to claim 8, wherein a determination result of either one of the results is selected. The claim that the learning unit learns the document data, the page number corresponding to the document data, the estimation result, the determination result of the first determination unit, and the determination result of the second determination unit. 8 or the image forming apparatus according to claim 9.

Images