Calibration support computer system and support method

Abstract

To provide support so as to properly calibrate a calibration target including an image and a character.SOLUTION: A calibration support computer system 3 includes: a calibration receiving unit 30 which receives data 20 to be calibrated from a use terminal 2; an extraction unit 31 which extracts an image portion 21 and a character portion 22 in the received data 20 to be calibrated; an image analysis unit 32 which analyzes accuracy r1 of the image portion 21 representing a predetermined image representation target X1; a character analysis unit 33 which analyzes accuracy r2 of the character portion 22 representing a predetermined character representation target X2; and a correlation analysis unit 34 which analyzes a degree of correlation R between the image representation target X1 and the character representation target X2. Results of the analysis are provided to the use terminal 2.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a proofreading support computer system and support method for supporting the proofreading of a proofreading object including an image and a character. [Background technology]

[0002] For example, merchandise related to a music group may be printed with image information such as illustrations or drawings of the members, as well as text information indicating the names the members call themselves (first name, nickname, alias, etc.), their dates of birth, hometown, and other attributes. When producing this type of printed matter, the client presents the contractor with design materials including an overview of the goods, a design concept, as well as photos or portraits of each person to be printed and descriptions of their attributes, and the contractor creates the design according to the information presented. The completed design is then proofread by printing a test version on paper.

[0003] Patent Document 1 discloses an input data re-editing device that extracts an image from input data for proofreading and enables editing using the original image. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6893712 Summary of the Invention [Problem to be solved by the invention]

[0005] Design creation and proofreading are done by humans, so mistakes are bound to happen. For example, the image or name of a person in a design may be mixed up, there may be a typo, or the font may be different from the specified one. Mistakes are especially likely to occur when there are multiple similar designs. And the proofreader may not notice such a mistake in the design. If a design with an overlooked error is printed, it may cause disadvantages to the client or customer. In view of the above circumstances, an object of the present invention is to provide support for accurate proofreading of a proofreading target that includes images and characters. [Means for solving the problem]

[0006] In order to solve the above problems, the present invention provides a proofreading support computer system for supporting proofreading of a proofreading object including an image and a character, comprising: a calibration receiving unit that receives calibration target data that has been digitized from the user terminal; an extraction unit that extracts image portions and character portions from the received proofreading target data; an image analysis unit for analyzing the likelihood that the image portion represents a predetermined image representation object; a character analysis unit that analyzes the likelihood that the character portion represents a predetermined character representation object; a correlation analysis unit that analyzes a degree of correlation between the image representation object and the character representation object; a providing unit that provides the analysis results by each of the analysis units to the user terminal; The present invention is characterized by comprising: The extraction unit and each analysis unit may include a trained model that has been machine-learned using artificial intelligence (AI) to perform the desired processing. The analysis results include an analysis result of the likelihood that the image portion represents a specified image representation object, an analysis result of the likelihood that the text portion represents a specified text representation object, and an analysis result of the degree of correlation between the image representation object and the text representation object.

[0007] The proofreading target data may include first surface data corresponding to a first surface of the proofreading target, and second surface data corresponding to a second surface of the proofreading target facing a side different from the first surface. In this case, it is preferable that the proofreading support computer system further includes an inter-surface analysis unit that analyzes an inter-surface correlation degree between the image representation object or the character representation object represented by the first surface data and the image representation object or the character representation object represented by the second surface data. This can assist in proofreading whether or not there is a correlation between the image or character on the first side and the image or character on the second side of the proofreading target. Either or both of the first and second face data may include the image portion. Either or both of the first and second face data may include the character portion.

[0008] the proofreading support computer system further comprises a label information receiving unit that receives label information of an object to be expressed from the user terminal; The image analysis unit analyzes a likelihood that the image portion represents an image representation target based on the label information, The character analysis unit may analyze a likelihood that the character portion represents a character representation target based on the label information. This can reduce the burden of identifying an image representation object from an image portion and the burden of identifying a character representation object from a character portion.

[0009] The method of the present invention is a method for supporting proofreading of a proofreading object including an image and a character by a proofreading support computer system, comprising the steps of: A step of receiving calibration target data obtained by digitizing the calibration target from a user terminal; extracting image and character portions from the received proofreading target data; analyzing the likelihood that the image portions represent a predetermined image object; analyzing the likelihood that the character portions represent a predetermined character representation object; analyzing a degree of correlation between the image representation object and the character representation object; providing each of the analysis results to the user terminal; The proofreading support computer system is characterized in that: Effect of the Invention

[0010] According to the present invention, it is possible to assist in accurately proofreading a proofreading target that includes images and characters. [Brief description of the drawings]

[0011] [Figure 1]Fig. 1(a) is a diagram showing an example of part of design creation information that is a reference for creating a design to be proofread in an embodiment of the present invention. Fig. 1(b) is a plan view showing an example of a created design. Fig. 1(c) is a plan view showing an example of a final design approved after proofreading the design in Fig. 1(b). [Diagram 2] FIG. 2 is an explanatory diagram showing an overview of the proofreading support system according to the first embodiment of the present invention in a state in which proofreading support is provided for the design in FIG. [Diagram 3] FIG. 3 is an explanatory diagram showing an example of a display screen of the analysis result of the proofreading support in the proofreading support system. [Figure 4] FIG. 4 is an explanatory diagram showing an overview of a proofreading support system according to a second embodiment of the present invention in a state in which proofreading support is provided for the design in FIG. [Diagram 5] FIG. 5 is an explanatory diagram showing an overview of a proofreading support system according to a third embodiment of the present invention in a state in which proofreading support is provided for the design in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. <First embodiment (FIGS. 1 to 3)> As shown in Fig. 1(b), the subject of proofreading in this embodiment is, for example, a design 10' for various goods of a music-related group. The design 10' may be printed on the goods themselves, or may be printed on the packaging of the goods. After the design 10' has been proofread one or more times, it becomes the final approved design 10, as shown in Fig. 1(c). Hereinafter, each piece of information in design 10' before and during proofreading will be indicated with a quotation mark (').

[0013] As shown in Fig. 1(a), a design 10' is created based on design creation information 19, etc., compiled for each person X (representation target) who is an individual member of each team in a group. As shown in Fig. 1, the design creation information 19, and thus the designs 10', 10, include image information 11, 11' in which each person X is an image representation target X1, and character (text) information 12, 12' in which each person X is a character representation target X2.

[0014] The image information 11, 11' may be illustrations 11a, 11a' or actual drawings of the person who is the image representation subject X1. The illustrations may be drawings of people in which the person is deformed into a two-headed figure or the like. The character information 12, 12' includes attribute information such as the nicknames 12a, 12a' (first name, nickname, stage name, etc.) of each person who is the character representation target X2, the team names 12b, 12b', the birth dates 12c, and the hometowns 12d. Furthermore, the character information 12, 12' may include non-attribute information separate from the attributes of each person, such as the event names and event dates 12e. The character information 12, 12' is not limited to being written in a general font, and may be designed or symbolized. The character information 12, 12' may include symbols.

[0015] As shown in FIG. 2, the proofreading support system 1 probabilistically analyzes what image information 11' and text information 12' in a design 10' represent (what attribute) of which person X (X1, X2), and uses this information for proofreading. In detail, the proofreading support system 1 is composed of a user terminal 2 that receives proofreading support, and a proofreading support computer system 3 that provides proofreading support information. The user terminal 2 is composed of a computer of a system user A, and is, for example, a smartphone, but is not limited to this and may be other computers such as a tablet, a notebook computer, or a desktop computer. Preferably, the user terminal 2 has a camera mechanism and a communication function. An example of system user A is a proofreader of the design 10'. The proofreader may be the person who orders the design 10', the person who accepts the order, or the designer.

[0016] 2, image data of a design 10' (a calibration target) photographed by the built-in camera 2c of the user terminal 2 becomes calibration target data 20. The calibration target data 20 includes an image portion 21 corresponding to the image information 11' in the design 10', and a character portion 22 corresponding to the character information 12' in the design 10'.

[0017] The image portion 21 includes an illustration 21a or actual drawing of a person who is the image representation subject X1. The character portion 22 includes character notations (text) expressing attribute information such as the nickname 22a of the person who is the character representation target X2, the name of the team he or she belongs to 22b, and non-attribute information such as the event name and the event date 22e. In this embodiment, for the sake of simplicity, it is assumed that only the attribute information of the attribute information and non-attribute information is the object of analysis.

[0018] As shown in FIG. 2, a user terminal 2 and a proofreading support computer system 3 are communicatively connected via a communication network 4 such as the Internet. The proofreading support computer system 3 is configured by a cloud server, but is not limited to this and may be configured by a physical server.

[0019] As shown in FIG. 2, the proofreading support computer system 3 functionally includes a proofreading receiving unit 30, an extraction unit 31, an image analysis unit 32, a character analysis unit 33, a correlation analysis unit , and a provision unit .

[0020] The proofreading receiving unit 30 receives, from the user terminal 2, photographed image data of the design 10', that is, proofreading target data 20 obtained by digitizing the proofreading target. The extraction unit 31 extracts the image portion 21 and the character portion 22 from the received proofreading target data 20. Preferably, the extraction unit 31 includes a trained model 31a constructed by machine learning using artificial intelligence so as to extract image portions and character portions from a large amount of learning photographic data.

[0021] The image analysis unit 32 analyzes the likelihood r1 that the image portion 21 represents a predetermined image representation target X1. For example, the image analysis unit 32 analyzes how close the image portion 21 is to a reference image of a person (image representation target X1) using the trained model 32a. Preferably, the trained model 32a is constructed to extract features from the analysis target image and the reference image of a person, etc., by metric learning using artificial intelligence, and calculate the similarity (r1). The reference image may be a photograph of the person, or may be a person illustration 11a represented in the design creation information 19.

[0022] The character analysis unit 33 analyzes the likelihood r2 that the character portion 22 represents a predetermined character representation target X2. For example, the character analysis unit 33 analyzes how similar the content written in the character portion 22 is to which attribute information of which person (character representation target X2) by using the trained model 33a. Preferably, the trained model 33a is constructed to extract features from the analysis target text and the reference text, such as person attributes, by metric learning using artificial intelligence, and calculate the similarity (r2). It is preferable that the trained model 33a is machine-learned so that it can analyze character fonts and different spellings (Roman letters, Katakana, Hiragana, etc.). Furthermore, it is preferable that the trained model 33a is machine-learned so that it can analyze special logos and designed characters. In analyzing the character portion 22, the accuracy may be analyzed for each item of attribute information (nickname 22a, team name 22b, etc.), or the accuracy may be calculated by integrating or averaging the accuracy of each item. The character portion relating to non-attribute information such as the event date 22e may also be subject to analysis by the character analysis unit 33.

[0023] The correlation analysis unit 34 analyzes the degree of correlation between the person (image representation target X1) represented by the image portion 21 analyzed by the image analysis unit 32 and the person attribute (character representation target X2) represented by the character portion 22 analyzed by the character analysis unit 33, that is, the correlation degree R (FIG. 3). In short, it analyzes the correlation between the image portion 21 and the character portion 22. Preferably, the correlation analysis unit 34 includes a trained model 34a constructed to analyze how close the image representation target X1 and the character representation target X2 are by metric learning using artificial intelligence. In analyzing the correlation degree, it is preferable to calculate the correlation degree with the image for each item of the character information 22 (nickname 22a, team name 22b, etc.). The contribution of the item of the character information 22 (nickname 22a, team name 22b, etc.) to the correlation degree may be weighted. For example, only the correlation degree between a specific item (e.g., nickname 22a) and the illustration 21a may be analyzed.

[0024] The providing unit 35 provides the analysis results by each of the analysis units 32, 33, 34 to the user terminal 2. As shown in Fig. 3, the analysis results include the accuracy r1 of the image representation object, the accuracy r2 of the text representation object, and the correlation degree R. As a form of provision, the providing unit 35 may transmit the analysis results to the user terminal 2, which may receive them. Alternatively, the providing unit 35 may upload the analysis results to a web page for user A of the system 1, which may be viewed on a dedicated or general-purpose browser of the user terminal 2.

[0025] The proofreading support system 1 is used, for example, as follows. System user A photographs a design 10' (object to be corrected) with the camera 2c of the user terminal 2. This generates correction object data 20 made up of the photographed image data. Proofreading target data 20 is uploaded to a proofreading support computer system 3 . In the proofreading support computer system 3, the proofreading receiving unit 30 receives the proofreading target data 20. Next, the extraction unit 31 extracts the image portion 21 and the character portion 22 from the received proofreading target data 20 .

[0026] Then, the image analysis unit 32 analyzes the likelihood r1 that the image portion 21 represents a predetermined person (image representation target X1). Specifically, the trained model 32a of the image analysis unit 32 identifies the person depicted in the illustration 21a of the image portion 21, and calculates the confidence level r1 of the identification. Furthermore, the character analysis unit 33 calculates the probability r2 that each of the items 22a and 22b in the character portion 22 represents a specific person (character representation target X2). Specifically, the trained model 33a of the character analysis unit 33 identifies the content written in each item in the character portion 22 as an attribute of the specific person, and calculates the confidence level r2 of the identification. Furthermore, the correlation analysis unit 34 analyzes the correlation degree R between the image representation object X1 and the character representation object X2. Thereafter, the providing unit 35 provides the analysis results X1, X2, r1, r2, and R from the analysis units 32, 33, and 34 to the user terminal 2.

[0027] 3, the accuracy r1, r2 and correlation R of the image portion 21 and the character portion 22 are displayed on the screen 2d of the user terminal 2. It is preferable to display the correlation r2 with the image calculated for each item (nickname 22a, team name 22b, etc.) of the character information 22 for each item. Furthermore, who (or what) the representation targets X1, X2 identified by the analysis units 32, 33 are may also be displayed on the viewing screen 2d.

[0028] System user A can accurately proofread design 10' by referring to the displayed contents. Since items with low accuracy r1 and r2 are likely to contain some kind of error, system user A can proofread with particular care. Furthermore, the degree of correlation R (match rate) can easily be used to determine the degree to which image information 11 (image) and character information 12 (text) are related. This can improve the accuracy of proofreading.

[0029] For example, as illustrated in Figure 1(b), when the person illustration 11a' in the design 10' is difficult to recognize as representing the person (image representation target X1), the accuracy r1 in the viewing screen 2d does not become a very high value, as shown in Figure 3. Furthermore, when the nickname 12a' or the team name 12b' in the design 10' is a typo or belongs to a different person, the accuracy r2 of the corresponding items 22a and 22b will have a low value. Furthermore, when the image information 11 (image) and the text information 12 (text) do not match, the correlation degree R takes a low value.

[0030] On the viewing screen 2d, items whose accuracy falls below a certain value may be highlighted by changing the size or color of the text, for example. On the viewing screen 2d, an item whose accuracy exceeds a certain value may be marked as "Passed" or the like. This allows a system user A, such as a proofreader, to easily find errors by paying attention to items in the design 10' that have low accuracy. The proofread results are presented to the designer. The designer then modifies the design 10' in accordance with the proofread results. In this manner, by carrying out proofreading and modification one or more times, an approved design 10 with the errors corrected is obtained, as shown in FIG. 1(c).

[0031] Next, another embodiment of the present invention will be described. In the following embodiments, the same reference numerals are given to the same aspects as those already described, and the description thereof will be omitted. <Second embodiment (FIG. 4)> As shown in FIG. 4, in the second embodiment of the present invention, design 10B' has design 13 on the front surface (first surface) and design 14 on the back surface (second surface facing the opposite side from the first surface).

[0032] The proofreading target data 20B of the proofreading support system 1B of the second embodiment includes first surface data 23 corresponding to the front surface design 13 and second surface data 24 corresponding to the back surface design 14. The proofreading target data 20B consisting of these surface data 23, 24 includes an image portion 21 and a character portion 22. Both the first surface data 23 and the second surface data 24 may include the image portion 21 and the character portion 22. One of the first surface data 23 and the second surface data 24 may include the image portion 21 and the character portion 22, and the other may include only the image portion 21 or only the character portion 22. One of the first surface data 23 and the second surface data 24 may include only the image portion 21, and the other may include only the character portion 22.

[0033] In the design 10B' and thus the proofreading target data 20B illustrated in FIG. 4, the first side data 23 includes an image portion 21 and a character portion 22, and the second side data 24 includes the character portion 22 but does not include an image portion. In FIG. 4, the first side data 23 and the second side data 24 are contained in one image file, but the first side data 23 and the second side data 24 may be contained in separate image files.

[0034] As shown in FIG. 4, a proofreading support computer system 3B of the second embodiment includes a surface identification unit 36 ​​and an inter-surface analysis unit 37. The surface identification unit 36 ​​distinguishes between the first surface data 23 and the second surface data 24 in the proofreading target data 20B received by the proofreading receiving unit 30. Preferably, the surface identification unit 36 ​​includes a trained model 36a constructed to distinguish between the first surface data 23 and the second surface data 24 by machine learning using artificial intelligence. Alternatively, the surface identification unit 36 ​​may automatically recognize an area on one side (e.g., the lower side) of the calibration target data 20B as first surface data 23, and automatically recognize an area on the other side (e.g., the upper side) of the calibration target data 20B as second surface data 24. When the first side data 23 and the second side data 24 are separate image files, the side identification unit 36 ​​may determine which image file is the first side data 23 and which image file is the second side data 24 from metadata and identification codes attached to these image files.

[0035] The extraction unit 31 extracts the image portion 21 and the character portion 22 contained in the first face data 23 and the second face data 24. Here, the image portion 21 and the character portion 22 are extracted from the first face data 23, and the character portion 22 is extracted from the second face data 24.

[0036] The method of analyzing the image portion 21 by the image analysis unit 32 and the method of analyzing the character portion 22 by the character analysis unit 33 are the same as those in the first embodiment. That is, the image analysis unit 32 analyzes the likelihood r1 that the image portion 21 of the first side data 23 and the second side data 24 each represent a predetermined image representation object. Here, the likelihood r1 that the image portion 21 of the first side data 23 represents a predetermined image representation object X1 is analyzed. Furthermore, the character analysis unit 33 analyzes the likelihood r2 that the character portions 22 of the first face data 23 and the second face data 24 each represent a predetermined character representation object X2.

[0037] 4, the inter-plane analysis unit 37 of the proofreading support computer system 3B analyzes the inter-plane correlation degree Rs. The inter-plane correlation degree Rs indicates the degree of correlation between an image representation object or a character representation object represented by the first plane data 23 and an image representation object or a character representation object represented by the second plane data 24. Preferably, the inter-plane analysis unit 37 includes a trained model 37a constructed to analyze the inter-plane correlation degree Rs by metric learning using artificial intelligence. This can help in calibrating whether or not the first surface 13 and the second surface 14 of the design 10B' to be calibrated are correlated.

[0038] <Third embodiment (FIG. 5)> As shown in Fig. 5, in a proofreading support system 1C of the third embodiment of the present invention, a label information input section 2b is provided on a menu screen for uploading proofreading target data of a user terminal 2. When proofreading target data 20 is uploaded to a proofreading support computer system 3C, label information 25 relating to the expression target of the proofreading target data 20 is input into the label information input section 2b. Examples of the label information 25 include a nickname 25a of a person who is the expression target X (see Fig. 1) and the name of the team to which the person belongs 25b. The label information 25 may be only the nickname 25a.

[0039] The label information 25 is sent to the proofreading support computer system 3C together with proofreading target data 20 (see FIG. 2) which is an image of the design 10'. The proofreading support computer system 3C of the third embodiment includes a label information receiving unit 38. The proofreading receiving unit 30 receives the proofreading target data 20 from the user terminal 2, and the label information receiving unit 38 receives the label information 25.

[0040] The image analysis unit 32 analyzes the likelihood r1 that the image portion 21 extracted from the proofreading target data 20 represents a person (image representation target X1) based on the label information 25. Specifically, the trained model 32a of the image analysis unit 32 analyzes how close the image portion 21 is to a reference image of a person corresponding to the name 25a, etc., indicated in the label information 25. Therefore, it is possible to narrow down the images to be compared and referenced with the image portion 21, and the process of identifying the image representation target X1 from the image portion 21 can be omitted or significantly simplified.

[0041] The character analysis unit 33 analyzes the likelihood r2 that the character portion 22 extracted from the proofreading target data 20 represents a person (character representation target X2) based on the label information 25. Specifically, the trained model 33a of the character analysis unit 33 analyzes how close the content written in the character portion 22 is to the attributes of a person corresponding to the nickname 25a, etc., indicated in the label information 25. This makes it possible to omit or greatly simplify the process of identifying the character representation target X2 from the character portion 22. If an incorrect input of the label information 25 occurs at the input stage at the user terminal 2, at least one of the degrees of accuracy r1, r2 and the degree of correlation R based on the label information 25 will be low. This allows the system user A to notice the incorrect input.

[0042] The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention. For example, the calibration target data is not limited to raster format image data created by photographing the calibration target with a camera, but may be vector data created by a computer application for creating designs, etc. The image portion 21 of the calibration target data 20 may include images of a plurality of people, and the image analysis section 32 may perform analysis for each person image.

[0043] The correlation analysis unit 34 may calculate the correlation R using a functional formula (1) in which the probability r1 that the image portion 21 represents a specified image representation object and the probability r2 that the character portion 22 represents a specified character representation object are variables (parameters). R = f(r1,r2) (1) The accuracy r2 in formula (1) may be the average value or the minimum value of the accuracy for each item of attribute information. The inter-surface analysis unit 37 of the second embodiment (Figure 4) may calculate the surface-surface correlation degree Rs according to a functional formula in which the probabilities r1, r2 for the first surface data 23 and the probabilities r1, r2 for the second surface data 23 are variables (parameters). The design 10' of the third embodiment (FIG. 5) is similar to that of the first embodiment (FIG. 2), but is not limited thereto and may include a first surface 13 and a second surface 14 as in the second embodiment (FIG. 4). The proofreading support computer system 3C of the third embodiment (FIG. 5) may have a surface identification unit 36 ​​and an inter-surface analysis unit 37 as in the second embodiment (FIG. 4).

[0044] The object of proofreading according to the present invention is not limited to designs of music group related goods, but can be applied to various proofreading objects as long as they include images and characters. [Industrial Applicability]

[0045] The present invention is applicable to a computer system that supports proofreading of designs for printed matter, for example. [Explanation of symbols]

[0046] A System User X Representation X1 Image expression target X2 Character expression target 1. Proofreading support system 1B,1C Proofreading Support System 2. Devices used 3. Proofreading Support Computer System 3B, 3C Calibration support computer system 10' Design (for proofreading) 10B' Design 11 Image information 12 Textual information 20 Data to be calibrated 20B Calibration target data 21 Image section 23 Front page data 24 Second page data 25 Label Information 30 Proofreading Department 31 Extraction part 32 Image Analysis Unit 33 Character analysis section 34 Correlation Analysis Department 35 Providing Department 36 Surface Identification Unit 37 Planar analysis section 38 Label Information Reception Department r1 The likelihood that an image part represents a given image representation object r2 The likelihood that a character part represents a given character representation object R correlation Rs Inter-surface correlation

Claims

1. A proofreading support computer system for supporting proofreading of a proofreading object including an image and a character, comprising: a calibration receiving unit that receives calibration target data that has been digitized from the user terminal; an extraction unit that extracts image portions and character portions from the received proofreading target data; an image analysis unit for analyzing the likelihood that the image portion represents a predetermined image representation object; a character analysis unit that analyzes the likelihood that the character portion represents a predetermined character representation object; a correlation analysis unit that analyzes a degree of correlation between the image representation object and the character representation object; a providing unit that provides the analysis results by each of the analysis units to the user terminal; A proofreading support computer system comprising:

2. the calibration object data includes first surface data corresponding to a first surface of the calibration object, and second surface data corresponding to a second surface of the calibration object facing a side different from the first surface, The proofreading support computer system of claim 1, further comprising an inter-plane analysis unit that analyzes the degree of inter-plane correlation between the image representation object or the character representation object represented by the first plane data and the image representation object or the character representation object represented by the second plane data.

3. A label information receiving unit that receives label information of an object to be expressed from the user terminal, The image analysis unit analyzes a likelihood that the image portion represents an image representation target based on the label information, The proofreading support computer system according to claim 1 , wherein the character analysis unit analyzes a likelihood that the character portion represents a character representation target based on the label information.

4. A method for supporting proofreading of a proofreading object including an image and a character by a proofreading support computer system, comprising: A step of receiving calibration target data obtained by digitizing the calibration target from a user terminal; extracting image and character portions from the received proofreading target data; analyzing the likelihood that the image portions represent a predetermined image object; analyzing the likelihood that the character portions represent a predetermined character representation object; analyzing a degree of correlation between the image representation object and the character representation object; providing each of the analysis results to the user terminal; The proofreading support method is characterized in that the proofreading support computer system executes the above steps.

Images