JP2020154951A - Font selection device and program - Google Patents

Abstract

To provide a font selection device and a program for reducing a burden of a layout editor.SOLUTION: A font selection device 1 includes: a learned model storage part 22 configured to store a learned model having learned, while associating with a font of a text, the text and format information on the text including arrangement information; and a font acquisition part 17 configured to acquire an inputted text and a font corresponding to the format information of the text, by using the learned model stored in the learned model storage part 22.SELECTED DRAWING: Figure 1

Description

The present invention relates to a font selection device and a program.

Conventionally, in the production process of magazines, books, etc., layout creation of the printing plate surface is generally performed manually by a layout editor using layout creation software regardless of whether the design is simple or complicated. Is going. For example, the layout of the print plate is newly created for each issue, which is a burdensome task for the layout editor.
As a device for solving this problem, for example, an editing device for generating a layout has been proposed (for example, Patent Document 1). In Patent Document 1, the layout target material is arranged based on the attribute information of whether the layout target material is automatically arranged or arranged at a fixed position and the design rule which is a rule for generating the layout. The decision is disclosed.

Japanese Unexamined Patent Publication No. 2008-147851

In creating the layout of the printing plate surface, the display mode of the text in which the font suitable for the desired impression is selected is also important for proper design. Further, in the layout of the printing plate surface, there are relatively many places having text. However, in the case described in Patent Document 1, the layout editor has to determine the font for each text and set the font for the text, which leaves a burden on the layout editor.

Therefore, an object of the present invention is to provide a font selection device and a program for reducing the burden on the layout editor.

The present invention solves the above-mentioned problems by the following solutions.
The first invention is a storage means for storing a learned model in which a text and format information related to the text including arrangement information are associated with a font of the text, and the input text and the text. It is a font selection device including a font acquisition means for acquiring a font corresponding to format information by using the trained model.
The second invention includes, in the font selection apparatus of the first invention, a text format acquisition means for acquiring text and format information of the acquired text from image data including character information, and the font acquisition means is the text. This is a font selection device that acquires a font corresponding to the text and the format information acquired by the format acquisition means using the trained model.
A third invention is a font selection device according to the second invention, which includes a font conversion means for converting character information of the image data into the font acquired by the font acquisition means.
A fourth invention is a font selection device in which the font acquired by the font acquisition means is a font image corresponding to the text in any of the font selection devices from the first invention to the third invention. is there.
A fifth invention is a font selection device according to any one of the first to third inventions, wherein the font acquired by the font acquisition means is code information corresponding to the text. is there.
According to a sixth aspect of the present invention, in any of the font selection devices from the first invention to the fifth invention, the font acquisition means acquires a plurality of the fonts for the text in association with a score indicating a recommendation order. , Font selection device.
A seventh invention is a font output means for outputting the text after applying a font obtained by the font acquisition means to the text in the font selection device of the sixth invention in association with the score. It is a font selection device provided.
The eighth invention is in any of the font selection devices from the first invention to the seventh invention, in which the format information includes at least the character size, character color, and background color of the text in addition to the arrangement information. It is a font selection device including one.
A ninth invention is a text information acquisition means for acquiring input text, the text format information, and the font of the text in any of the font selection devices from the first invention to the eighth invention. A font selection device including the text acquired by the text information acquisition means, the format information, and a learning means for constructing the trained model using the font as teacher data.
The tenth invention is a program for operating a computer as any font selection device from the first invention to the ninth invention.

According to the present invention, it is possible to provide a font selection device and a program for reducing the burden on the layout editor.

It is an overall outline diagram of the font selection system which concerns on this embodiment, and the functional block diagram of the font selection apparatus. It is a flowchart which shows the model construction process of the font selection apparatus which concerns on this embodiment. It is a flowchart which shows the font acquisition process of the font selection apparatus which concerns on this embodiment. It is a figure for demonstrating the font acquisition process which concerns on this embodiment with a specific example. It is a figure which shows the processing example by the font acquisition processing which concerns on this embodiment. It is a figure which shows the processing example by the font acquisition processing which concerns on this embodiment.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. It should be noted that this is only an example, and the technical scope of the present invention is not limited to this.
(Embodiment)
<Font selection system 100>
FIG. 1 is an overall schematic diagram of the font selection system 100 according to the present embodiment and a functional block diagram of the font selection device 1.

The font selection system 100 shown in FIG. 1 is a system including a font selection device 1 and a terminal 4. The font selection device 1 receives information (image data) about the layout being created including the text to be selected as the font from the terminal 4, and acquires the text and the format information about the text from the received information about the layout. .. Then, the font selection device 1 takes the acquired text and the format information related to the text as input, and acquires the font for the text using the trained model. After that, the font selection device 1 presents the acquired font to the terminal 4.
The font selection device 1 and the terminal 4 are communicably connected via the communication network N. The communication network N is, for example, an Internet line or the like, regardless of whether it is wired or wireless.

<Font selection device 1>
The font selection device 1 is a device that selects a font from text and format information related to the text by using a trained model for selecting a font.
Further, the font selection device 1 is a device for constructing a trained model using the text, the format information related to the text, and the font as teacher data.

The font selection device 1 is, for example, a server. The font selection device 1 may also be a personal computer (PC) or the like.
As shown in FIG. 1, the font selection device 1 includes a control unit 10, a storage unit 20, and a communication interface unit 29.
The control unit 10 is a central processing unit (CPU) that controls the entire font selection device 1. The control unit 10 appropriately reads and executes the operating system (OS) and the application program stored in the storage unit 20 to cooperate with the above-mentioned hardware and execute various functions.

The control unit 10 includes a model construction unit 11 and a font selection unit 15.
The model building unit 11 builds a trained model for selecting a font.
The model building unit 11 includes a text information acquisition unit 12 (text information acquisition means) and a learning unit 13 (learning means).

The text information acquisition unit 12 uses, for example, text, format information, and fonts as teacher data from the created layout in which the layout has been created by using layout creation software or the like and the font is manually set. get. Here, the created layout constitutes, for example, a printing plate surface. Further, the printed plate surface may be a cover of a magazine or a book, or may be a page of a magazine or the like. The effect of the present invention can be further exhibited when a large number of texts are arranged at various positions on the printing plate surface and include images such as photographs and illustrations.

The text is each string contained in the created layout. The text may be, for example, a word or a cohesive document. In the case of a cohesive document, for example, one cohesive until reaching a blank or a punctuation mark may be a text.
The format information is information about text, and includes, for example, information such as an arrangement position which is a layout of text, a size of text, and a color of text. The arrangement position may be the coordinate value itself. In addition, the arrangement position is not the coordinate value itself, but an area is defined in advance in the layout (for example, the upper left is the coordinate (0,0), and "(0,0)-(30,30): upper row". Etc.), it is also possible to specify an area within that range. The size of the text can be represented by, for example, points (pt). The color of the text may be, for example, "red", "black", or may be represented by a CMYK (Cyan-Magenta-Yellow-blacK) value.
The font is information representing typeface data such as Mincho and Gothic. The font may have a name such as Mincho. Further, the font may be an image of a typeface originally created, and in that case, the font may be identification information representing an image of a unique typeface.

Here, the text information acquisition unit 12 can acquire teacher data from various data possessed by the layout creation software, for example, when the layout is generated by using the layout creation software.
However, there are cases where there is no information converted into data, such as pages of paper magazines, but it is desired to be a learning target. In such a case, the text information acquisition unit 12 can obtain text from the image of the generated layout by using the function of OCR (Optical character recognition) such as an image scanner (not shown). In addition, the text information acquisition unit 12 can obtain format information by a known image processing technique. Further, the text information acquisition unit 12 can estimate the font of the text by using, for example, a font classifier (not shown) which has been trained in advance. Then, the text information acquisition unit 12 can use various information obtained by the above method as teacher data.
The text information acquisition unit 12 may be acquired by the user who constructs the trained model by manually setting (inputting) the text, the format information, and the font and accepting them as teacher data.

The learning unit 13 executes machine learning using the teacher data acquired by the text information acquisition unit 12 as input data, and generates a trained model. As an algorithm for machine learning, in this embodiment, for example, supervised learning is performed by a recurrent neural network (RNN). Here, the reason why the recursive neural network is used is to regard the text as time series data and to consider the order of appearance of characters. By doing so, when the text is a meaningful word or document, learning can be performed including the sequence of the word or document. Then, the learning unit 13 stores the generated learned model in the trained model storage unit 22.

Since machine learning by a recursive neural network is a known technique, detailed description thereof will be omitted. Further, as the machine learning algorithm, for example, another algorithm such as a convolutional neural network (CNN) may be used as long as the text is not regarded as time series data.

The font selection unit 15 uses the text and the format information as input data, and acquires a font using the trained model constructed by the model construction unit 11.
The font selection unit 15 includes a text format acquisition unit 16 (text format acquisition means), a font acquisition unit 17 (font acquisition means), and a layout update unit 18 (font conversion means).

The text format acquisition unit 16 acquires text and format information from the layout being created that has generated a layout other than the font. For example, when the layout is generated by using the layout creation software, the layout creation software has various data such as text, text arrangement position, character size, and character color. In that case, the text format acquisition unit 16 may acquire text and format information from various data included in the layout creation software. Further, the text format acquisition unit 16 may acquire the data by manually inputting the text and the format information by the layout editor.

The font acquisition unit 17 applies the text acquired by the text format acquisition unit 16 and the format information to the trained model as input of the trained model, and acquires the font as the output of the trained model.
The layout update unit 18 converts the layout text using the font acquired by the font acquisition unit 17.
The details of each functional unit will be described later.

The storage unit 20 is a storage area for a hard disk, a semiconductor memory element, or the like for storing programs, data, and the like necessary for the control unit 10 to execute various processes.
The storage unit 20 includes a program storage unit 21 and a learned model storage unit 22 (storage means).
The program storage unit 21 is a storage area for storing various programs. The program storage unit 21 stores the font selection program 21a. The font selection program 21a is a program for performing various functions executed by the control unit 10 of the font selection device 1. In addition to the font selection program 21a, the storage unit 20 may include, for example, a program such as layout creation software that automatically generates a layout.

The trained model storage unit 22 is a storage area for storing the trained model constructed by using the teacher data.
The communication interface unit 29 is an interface for performing communication with, for example, the terminal 4 via the communication network N.

<Terminal 4>
The terminal 4 shown in FIG. 1 is, for example, a PC used by a layout editor or the like. The terminal 4 may also be a portable terminal having a computer function represented by a tablet.
The terminal 4 transmits, for example, the layout being created to the font selection device 1, and receives the layout in which the font has been changed from the font selection device 1.
Although not shown, the terminal 4 includes a control unit, a storage unit, an input unit, a display unit, a communication interface unit, and the like.

Here, the computer refers to an information processing device provided with a control unit, a storage device, etc., and the font selection device 1 and the terminal 4 are information processing devices provided with a control unit, a storage unit, etc., respectively, and are a concept of a computer. include.

<Processing of font selection system 100>
Next, the processing in the font selection system 100 will be described.
First, the process of constructing the trained model will be described with reference to FIG.
FIG. 2 is a flowchart showing a model construction process of the font selection device 1 according to the present embodiment.

In step S 11 of FIG. 2 (hereinafter, simply referred to as “S”) 11, for example, the control unit of the terminal 4 transmits the created layout to the font selection device 1, thereby controlling the font selection device 1. The unit 10 (model construction unit 11) accepts the created layout.
In S12, the control unit 10 (text information acquisition unit 12) acquires text, format information, and a font from the received created layout, and uses these as teacher data.
In S13, the control unit 10 (learning unit 13) performs supervised learning by a recursive neural network as machine learning.
In S14, the control unit 10 (learning unit 13) stores the learned model constructed by machine learning in the learned model storage unit 22. After that, the control unit 10 ends this process.

In the model construction process described above, it is desirable to perform machine learning using teacher data acquired from a large number of created layouts as input data.
Further, in the model building process, it is desirable to build a trained model for each category of publishers and magazines, for example. By doing so, it will be possible to create a consistent design layout for each publisher and magazine.
Further, the trained model may be constructed for each character size and arrangement. By doing so, it is possible to build a trained model specialized in character size and arrangement according to the application.

Next, the process of selecting a font using the constructed trained model will be described with reference to FIG.
FIG. 3 is a flowchart showing a font acquisition process of the font selection device 1 according to the present embodiment.
For example, when a program related to font acquisition processing is plugged into layout creation software, a layout editor or the like performs font conversion while the layout being created using the layout creation software is displayed on the terminal 4. Give instructions. By doing so, the layout (image data including character information) being created from the terminal 4 is transmitted to the font selection device 1, and this process is started.

In S21 of FIG. 3, the control unit 10 (text format acquisition unit 16) acquires text and format information from the layout being created.
In S22, the control unit 10 (font acquisition unit 17) applies the text acquired in S21 to the trained model and the format information to acquire the font.

Here, when the text and the format information are input information to the trained model, the font output means of the control unit 10 outputs a plurality of types of fonts from the trained model and orders the fonts. You may. Specifically, the font output means of the control unit 10 may output the font output from the trained model to the terminal 4 together with the score with a score indicating the degree of application. Then, the terminal 4 may display the font together with the score, and when the layout editor or the like selects the font, the selected font may be transmitted to the font selection device 1.
In S23, the control unit 10 (layout update unit 18) updates the layout to which the acquired font is applied. That is, the control unit 10 converts the text acquired from the layout being created into the font acquired in S22. The updated layout is transmitted to, for example, the terminal 4, and the terminal 4 displays the layout in which the font is changed. After that, the control unit 10 ends this process.

Here, font selection will be described based on a specific example.
FIG. 4 is a diagram for explaining the font acquisition process according to the present embodiment together with a specific example.
First, when the creating layout 51 of the font selection target is input to the font selection device 1, the control unit 10 obtains the input data 61 from the creating layout 51 (S21 in FIG. 3). The input data 61 includes a character string "good morning" which is a text. Further, the input data 61 includes the arrangement position which is the format information is "upper center", the character size is "12pt", and the character color is "black". Here, the background color and the like may be included as the format information.
Next, the control unit 10 obtains the output data 62 by applying the input data to the trained model stored in the trained model storage unit 22 (S22 in FIG. 3). The output data 62 includes, for example, the name of the font, "HG Soei Kaku Pop Body".
Then, the control unit 10 generates a layout 71 in which the font of the output data 62 is reflected in the text (S23 in FIG. 3). The layout 71 has a different font for the text of "Good morning" as compared with the layout 51 being created.

Next, as an algorithm for machine learning, an example of a font that can be generated by a learning model constructed using a recursive neural network is shown.
FIG. 5 is a diagram showing a processing example by the font acquisition processing according to the present embodiment.
In the layout 53 being created shown in FIG. 5, the text is "Recommended Nara Sanpo". This text is a combination of the words "recommended", "nara", and "sanpo".
The layout 73 reflects the font obtained by applying the input data including the text to the trained model stored in the trained model storage unit 22 in the text. Here, the trained model is constructed by using a recursive neural network as described above. Therefore, the texts can be read in order, and for example, meaningful words can be made into different fonts. In the case of this example, the trained model has learned about the printed side of the tourist guidebook, and therefore the font with the place name "Nara" emphasized is selected.

Next, an example of the cover of a magazine is shown.
FIG. 6 is a diagram showing a processing example by the font acquisition processing according to the present embodiment.
The layout 55 under construction shown in FIG. 6 relates to the cover of a magazine. Then, of the layout 55 being created, the font of the text 55a indicating "Nara" is converted by using the learning model.
The layout 75 includes the text 75a obtained by reflecting the font obtained by applying the input data including the text 55a to the trained model stored in the trained model storage unit 22 in the text 55a.
As described above, the font selection device 1 outputs the text 55a included in the layout 55 being created as the layout 75 including the text 75a after the font is converted. Therefore, it is possible to generate a layout of a suitable font while creating the layout, which improves convenience.

As described above, the font selection device 1 of the present embodiment has the following effects.
(1) Using the trained model learned using the text, the format information, and the teacher data including the font, the text and the format information are used as input data, and the font is acquired. Therefore, for example, when it is used for the layout being created, the font suitable for the layout is acquired, so that the burden on the layout editor can be reduced.
(2) Since the text and format information are acquired from the layout being created, various data used as input data of the trained model can be easily acquired, especially when using layout creation software. it can.
(3) Since the acquired font is applied to the corresponding text of the layout being created and the layout is updated to the layout in which the font is converted, the convenience for creating the layout including the font is improved.

(4) When a font image is used as teacher data when the trained model is generated, the font acquired by the font acquisition process becomes the font image. Therefore, instead of a general font, a font of a typeface such as a decorative character can be used.
(5) When the code information indicating the font is used as the teacher data when the trained model is generated, the acquired font becomes the code information indicating the font. Therefore, in the trained model, it is sufficient to have the code information, and the amount of data related to the communication between the font selection device 1 and the terminal 4 can be suppressed.

(6) Since a plurality of fonts are acquired in association with the score indicating the recommended order and the font applied to the text is output, the layout editor can select from the plurality of fonts. In addition, the score can be used as a reference when making a selection.
(7) Since the format information includes at least one of the character size, character color, and background color of the text in addition to the arrangement information, the font of the text can be made more suitable for the layout.
(8) For example, it is also possible to acquire the text input based on the completed layout, the format information, and the font, and construct a trained model using each acquired data as teacher data. Therefore, the constructed trained model can be made to reflect the intentions of the actual layout editor and the like, and can be made highly complete.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Moreover, the effects described in the embodiments are merely a list of the most preferable effects arising from the present invention, and the effects according to the present invention are not limited to those described in the embodiments. The above-described embodiment and the modified form described later can be used in combination as appropriate, but detailed description thereof will be omitted.

(Transformed form)
(1) In the present embodiment, an example of accessing the font selection device 1 from the terminal 4 and performing processing has been described, but the present invention is not limited to this. The font selection device may be a stand-alone personal computer (PC) or the like having an input unit and a display unit.

(2) In the present embodiment, the construction of the trained model and the selection of fonts have been described by taking as an example a font selection device, but the present invention is not limited to this. The construction of the trained model and the selection of the font may be separate devices. In that case, the trained model generated by the device for constructing the trained model is transplanted to the device for selecting the font, or the device for selecting the font and the trained model are constructed to have the trained model. The device and the device for selecting the font may be connected so as to be able to communicate with each other.
Further, in the present embodiment, the font selection program has been described as performing both the process of constructing the trained model and the process of selecting the font, but the present invention is not limited to this. The program for constructing the trained model and the program for selecting the font may be different programs.

(3) In the present embodiment, the font acquisition process has been described as a process in the case of a plug-in of layout creation software, but the present invention is not limited to this. Of course, it may be performed separately from the layout creation software.
(4) In the present embodiment, it has been described that the format information may include at least one of the character size, the character color, and the background color of the text in addition to the arrangement information, but the format information is not limited to this. It may be the ratio of the text area to the total size of the layout.

1 Font selection device 4 Terminal 10 Control unit 11 Model construction unit 12 Text information acquisition unit 13 Learning unit 15 Font selection unit 16 Text format acquisition unit 17 Font acquisition unit 18 Layout update unit 20 Storage unit 21a Font selection program 22 Learned model storage Part 100 Font selection system

Claims (10)

A storage means for storing a learned model in which a text and format information related to the text including arrangement information are associated with a font of the text and learned.
A font acquisition means for acquiring a font corresponding to the input text and the format information of the text by using the trained model, and
Font selection device equipped with. In the font selection device according to claim 1,
A text format acquisition means for acquiring text and acquired format information of the text from image data including character information is provided.
The font acquisition means is a font selection device that acquires fonts corresponding to the text and the format information acquired by the text format acquisition means using the trained model. In the font selection device according to claim 2.
A font selection device including a font conversion means for converting character information of the image data into the font acquired by the font acquisition means. In the font selection device according to any one of claims 1 to 3.
The font selected by the font acquisition means is a font image corresponding to the text. In the font selection device according to any one of claims 1 to 3.
The font acquired by the font acquisition means is a font selection device which is code information corresponding to the text. In the font selection device according to any one of claims 1 to 5.
The font acquisition means is a font selection device that acquires a plurality of the fonts for the text in association with scores indicating a recommendation order. In the font selection device according to claim 6,
A font selection device comprising a font output means for outputting a text after applying a font obtained by applying the font acquired by the font acquisition means to the text in association with the score. In the font selection device according to any one of claims 1 to 7.
The format information is a font selection device that includes at least one of a character size, a character color, and a background color of the text in addition to the arrangement information. In the font selection device according to any one of claims 1 to 8.
A text information acquisition means for acquiring the input text, the format information of the text, and the font of the text,
A learning means for constructing the trained model using the text acquired by the text information acquisition means, the format information, and the font as teacher data.
Font selection device equipped with. A program for operating a computer as the font selection device according to any one of claims 1 to 9.

Images