Information processing device, writing system, information processing program, and information processing method

Abstract

To provide an information processing device which is configured to create a block copy using in combination a handwriting-style font dynamically created and a handwriting-style font preliminarily stored in memory means.SOLUTION: In an information processing device, reception means receives a character code, classification means classifies the character code received by the reception means into a predetermined first character and a predetermined second character, font creation means dynamically creates a handwriting-style font for the first character, call means calls a handwriting-style font stored in memory means for the second character, and block copy creation means creates a block copy using the font created by the font creation means and the font called by the call means.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing device, a writing system, an information processing program, and an information processing method. [Background technology]

[0002] There is a technology that generates characters. As a related technology, for example, Patent Document 1 discloses a character pattern generation method as the first vector-based character generation technology, in which the character pattern of each character is composed of a plurality of straight lines, and each straight line portion is treated as a vector, and the number of vectors required to represent one character, the starting point, direction, and length of each vector are stored as information in a main memory, and a character pattern to be displayed is selected from the memory, read out as information for each vector, and these are written sequentially into a random access memory in correspondence with an arbitrarily set coordinate system, and the information written into the random access memory can be read out by appropriate reproduction means.The applicant (inventor) of this application is the inventor of Patent Document 1, and is also the inventor of U.S. Patent No. 3,936,664.

[0003] Furthermore, for example, Patent Document 2 discloses a configuration in which, with the aim of improving the speed at which stroke characters are generated, a main processor specifies the stroke characters to be displayed using a code, and a stroke character generator generates stroke characters in accordance with this code; when generating stroke characters, the processor of the stroke character generator outputs constant component values ​​for a predetermined number of vectors that make up the stroke characters, and a dot generator of the stroke character generator inputs these constants and sequentially generates coordinate values ​​for a predetermined number of dots that make up the vectors; the processor repeatedly performs this operation for all vectors that make up the stroke characters; and since there is no need to calculate the component values ​​for each vector, the speed at which stroke characters are generated is significantly improved.

[0004] There is also known a technology for generating fonts that reproduce handwritten characters. For example, Patent Document 3 discloses a terminal device that compares a handwritten character image with a base font image to extract distortion data, and then expands the distortion data into multiple characters of the base font to generate handwritten-style font data.

[0005] Furthermore, for example, Patent Document 4 discloses that the object of the character generation device is to provide a character generation device that can reproduce handwritten characters while taking into account intra-individual variation, and the character generation device comprises a storage unit that stores a model font consisting of a group of character data corresponding to a plurality of character codes, a character data acquisition unit that acquires character data, a character identification unit that identifies the character corresponding to the acquired character data, a deformation amount extraction unit that compares the acquired character data with character data of the model font that corresponds to the character identified by the character identification unit, and extracts a deformation amount from the difference value of the feature amount, and a character data generation unit that generates new character data by associating it with the corresponding character code using a value obtained by adding or subtracting the deformation amount for each character data of the model font.

[0006] Furthermore, for example, Non-Patent Document 1 describes a computer generating handwritten characters using a recurrent neural network (RNN), which is a type of neural network. The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Special Publication No. 54-7416 [Patent Document 2] Japanese Patent Application Publication No. 5-249946 [Patent Document 3] Patent No. 5834822 [Patent Document 4] Japanese Patent Application Publication No. 2019-28094 [Non-patent literature]

[0008] [Non-Patent Document 1] Alex Graves, Department of Computer Science, University of Toronto, "Generating Sequences With Recurrent Neural Networks," [online], 5 Jun 2014, Internet〈URL:https: / / arxiv.org / abs / 1308.0850〉 Summary of the Invention [Problem to be solved by the invention]

[0009] There is a demand for characters that look handwritten, as opposed to printed characters, that is, characters that are close to the shape of characters written by hand. To create the feel of handwriting, the character shape does not remain fixed, but rather changes each time the same character is written multiple times, which is a characteristic of handwritten writing. As a prior art that generates different character shapes each time, there is an example of generating alphanumeric characters by applying a recurrent neural network, as in the aforementioned Non-Patent Document 1. However, this only covers a maximum of 256 alphanumeric characters, and implementing Japanese, which requires at least 3,000 characters for text expression, is not possible with the processing power of a typical PC, due to the amount of data required for machine learning and the computer processing power required to achieve it. For this reason, it is currently difficult to commercially implement handwritten-style characters in Japanese. Japanese is an example, and the same applies to other languages ​​(e.g., Chinese, etc.). Therefore, the present invention aims to provide an information processing device, writing system, information processing program, and information processing method that generate a copy of a worksheet by combining dynamically generated handwritten-style fonts and handwritten fonts that are pre-stored in a storage means. [Means for solving the problem]

[0010] The gist of the present invention to achieve this object resides in the following inventions. Invention [1] is an information processing device having a receiving means for receiving a character code, a classification means for classifying the character code received by the receiving means into predetermined first characters and second characters, a font generation means for dynamically generating a handwritten-style font for the first characters, a calling means for calling a handwritten font stored in a storage means for the second characters, and a copy generation means for generating a copy using the font generated by the font generation means and the font called by the calling means.

[0011] Invention [2] is an information processing device according to invention [1], wherein the font generation means has a font generation means for each character type, and dynamically generates a font using a font generation means corresponding to the character type of the first character.

[0012] Invention [3] is an information processing device according to invention [1], in which the copy generation means assigns information indicating the order of characters to be written to the font generated by the font generation means and the font called by the calling means in accordance with the order of appearance of character codes in the text accepted by the accepting means.

[0013] Invention [4] is an information processing device according to invention [1], further comprising a conversion means for performing an affine transformation or a homography transformation on the font called by the calling means, and the copy-over generation means generates a copy-over using the font generated by the font generation means and the font converted by the conversion means.

[0014] Invention [5] is the information processing device according to invention [4], wherein the coefficients in the affine transformation or the homography transformation in the transformation means are changed by a random function.

[0015] Invention [6] is an information processing device according to invention [5], in which the timing for changing the coefficient to a value fluctuating using a random function is either every time a font is generated or when the same character code appears in the layout.

[0016] Invention [7] is an information processing device according to invention [1], which rotates the font of alphanumeric character codes by 90 degrees clockwise when writing vertically.

[0017] Invention [8] is a writing system having an information processing device and a writing device described in any one of Inventions [1] to [7], wherein the font is described in vector format, and the writing device has an output means for writing characters according to the vector representing the font in the copy generated by the copy generation means.

[0018] Invention [9] is an information processing program that causes a computer to function as: a receiving means for receiving a character code; a classification means for classifying the character code received by the receiving means into predetermined first characters and second characters; a font generation means for dynamically generating a handwritten-style font for the first characters; a calling means for calling a handwritten font stored in a storage means for the second characters; and a copy generation means for generating a copy using the font generated by the font generation means and the font called by the calling means.

[0019] Invention

[10] is an information processing method performed by an information processing device that can access a storage means, wherein the information processing device performs the following steps: a reception step for receiving a character code; a classification step for classifying the character code received by the reception step into predetermined first and second characters; a font generation step for dynamically generating a handwritten-style font for the first character; a call step for calling a handwritten font stored in the storage means for the second character; and a plate generation step for generating a plate using the font generated by the font generation step and the font called by the call step. [Effects of the Invention]

[0020] According to the information processing device described in the invention [1], it is possible to generate a layout copy by mixing a dynamically generated handwritten font and a handwritten font stored in advance in a storage means.

[0021] According to the information processing device described in the invention [2], fonts can be dynamically generated using font generation means corresponding to the character type.

[0022] According to the information processing device described in the invention [3], it is possible to generate a layout that does not require a one-to-one correspondence between character codes and fonts.

[0023] According to the information processing device described in the invention [4], it is possible to impart fluctuations to handwritten fonts stored in advance in a storage means.

[0024] According to the information processing device described in the invention [5], it is possible to provide fluctuations to the coefficients in the affine transformation or homography transformation.

[0025] According to the information processing device described in the invention [6], even for handwritten style fonts stored in advance in a storage means, when the same character is written multiple times, the character shape can be changed each time.

[0026] According to the information processing device described in invention [7], when writing vertically, an alphanumeric font rotated 90 degrees clockwise can be used.

[0027] According to the writing system described in invention [8], characters can be written according to vectors that indicate fonts.

[0028] According to the information processing program described in the invention [9], a layout copy can be generated by combining a dynamically generated handwritten font and a handwritten font stored in advance in a storage means.

[0029] According to the information processing method described in invention

[10] , a dynamically generated handwritten font and a handwritten font stored in advance in a storage means can be mixed to generate a layout copy. [Brief explanation of the drawings]

[0030] [Figure 1] FIG. 2 is a conceptual module configuration diagram of a configuration example of the present embodiment. [Figure 2] FIG. 1 is an explanatory diagram showing a schematic configuration example using the present embodiment. [Figure 3] FIG. 10 is an explanatory diagram showing a more specific configuration example using the present embodiment. [Figure 4] FIG. 1 is an explanatory diagram showing an example of the configuration of a system using the present embodiment. [Figure 5] FIG. 10 is an explanatory diagram showing an example of writing according to the present embodiment. [Figure 6] 10 is a flowchart illustrating an example of processing according to the present embodiment. [Figure 7] FIG. 10 is an explanatory diagram showing a processing example according to the present embodiment. [Figure 8] FIG. 10 is an explanatory diagram showing a processing example according to the present embodiment. [Figure 9] FIG. 10 is an explanatory diagram showing a processing example according to the present embodiment. [Figure 10] FIG. 2 is a block diagram showing an example of the hardware configuration of a computer that realizes the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0031] Hereinafter, a preferred embodiment for realizing the present invention will be described with reference to the drawings. FIG. 1 shows a conceptual module configuration diagram of an example of the configuration of this embodiment. A module generally refers to a logically separable piece of software (including a computer program when interpreting "software"), hardware, or other component. Therefore, in the present embodiment, a module refers not only to a module in a computer program but also to a module in a hardware configuration. Therefore, the present embodiment also describes computer programs (e.g., programs for causing a computer to execute respective procedures, programs for causing a computer to function as respective means, and programs for causing a computer to realize respective functions) and systems and methods for functioning as those modules. For convenience of explanation, however, terms such as "store" and "cause to store" are used. However, when the embodiment is a computer program, these terms mean storing in a storage device or controlling to store in a storage device. Furthermore, modules may correspond one-to-one to functions, but in implementation, one module may be composed of one program, multiple modules may be composed of one program, or conversely, one module may be composed of multiple programs. Furthermore, multiple modules may be executed by a single computer, or one module may be executed by multiple computers in a distributed or parallel environment. Furthermore, one module may contain other modules. Furthermore, hereinafter, the term "connection" refers not only to a physical connection but also to a logical connection (e.g., data transmission / reception, instructions, reference relationships between data, login, etc.). "Predetermined" refers to being determined before the target processing, and includes being determined according to the situation or state at the time, or according to the situation or state up to that point, not only before the processing according to this embodiment begins, but also after the processing according to this embodiment begins, as long as it is before the target processing. When there are multiple "predetermined values," they may be different values, or two or more values ​​("two or more values" includes, of course, all values) may be the same. Furthermore, the phrase "if A, do B" is used to mean "determine whether A is true, and if it is determined that A is true, do B."However, this does not include cases where it is not necessary to determine whether it is A or not. In addition, when things are listed as "A, B, C," etc., this is an illustrative list unless otherwise specified, and includes cases where only one of them is selected (for example, only A). Furthermore, a system or device may be composed of multiple computers, hardware, devices, etc. connected by a communication means such as a network ("network" includes one-to-one communication connections), but it also includes cases where it is realized by a single computer, hardware, device, etc. "Device" and "system" are used as synonyms. Of course, a "system" does not include anything that is merely a social "mechanism" (i.e., a social system) that is an artificial arrangement. Furthermore, for each process by each module, or for each process when multiple processes are performed within a module, the target information is read from the storage device, and after the process is performed, the processing results are written to the storage device. Therefore, explanations of reading from the storage device before processing and writing to the storage device after processing may be omitted. Of course, the purpose of providing an object (device, etc.), method, or program relating to the embodiments described below, or the purpose of providing an object (device, etc.), method, or program relating to the invention that can be understood from the embodiments, may also be considered as the "problem that the invention aims to solve."

[0032] The problem that the present embodiment aims to solve may be understood as follows. Although the present embodiment is a concrete embodiment of the "problem that the invention aims to solve," it goes without saying that the present embodiment solves the "problem that the invention aims to solve." This explanation (the description in this paragraph number) is intended to facilitate understanding of the present embodiment, and is not intended to be used for restrictive interpretation. It goes without saying that this explanation alone should not be used to determine whether the invention for which a patent is sought is one described in the detailed description of the invention (Article 36, Paragraph 6, Item 1 of the Patent Act). Output systems using standard fonts only produce flat typefaces that fail to attract the recipient's attention.Handwritten letters, etc., have a high open rate and are easy to read due to their novelty, making them particularly effective for direct sales mail. In other words, there is a demand for characters that look handwritten, as opposed to printed characters, that is, characters that are close to the shapes that are written by hand by a person. Furthermore, even in the handwritten character generation devices disclosed in the aforementioned Patent Documents 3 and 4, which cannot be produced using normal fonts, the same characters have the same shape, which is unnatural, because the fonts are fixed, and a character generation technology that gives a more handwritten feel is required. To create the feel of handwriting, the character shape does not remain fixed, but rather changes each time the same character is written multiple times, which is a characteristic of handwritten writing. As a prior art that generates different character shapes each time, there is an example of generating alphanumeric characters by applying a recurrent neural network, as in the aforementioned Non-Patent Document 1. However, this only covers a maximum of 256 alphanumeric characters. Implementing Japanese, which requires at least 3,000 characters for text expression, is impossible with the processing power of a typical PC, given the amount of data required for machine learning and the computer processing power required to achieve it. For this reason, creating handwritten-style characters is currently not commercially feasible for Japanese. Specifically, in Japanese, personal and place names require more characters than the common kanji (approximately 2,100 characters). In other words, the Basic Resident Register characters (approximately 19,000 characters) and the Family Register Unified Characters (55,270 kanji characters) are used, and it is necessary to be able to create and use character forms for these as needed. Japanese is an example; the same applies to other languages ​​(e.g., Chinese). Although it is possible to prepare multiple font sets and select fonts at random, preparing a large number of font sets with handwritten designs is difficult in terms of time and cost in terms of the number of man-hours required. Therefore, the present embodiment aims to provide an information processing device, writing system, information processing program, and information processing method that combines dynamically generated handwritten-style fonts with handwritten fonts stored in advance in a storage means to generate a copy that gives the overall impression that the characters have fluctuations each time they are written.

[0033] In this embodiment, all characters including common kanji, basic resident register characters, and unified family register characters can be written. In other words, all of these characters can be stored in the storage module 150 described below. However, storing all of these characters in the storage module 150 is not an essential constituent requirement of the invention. In other words, it is sufficient to store the characters required for writing in the storage module 150 as a "predetermined font."

[0034] The information processing device 100 according to this embodiment has a function of generating a layout using a handwritten font. As shown in the example of Fig. 1, the information processing device 100 has a reception module 110, a classification module 120, a font generation module 130, a call module 140, a storage module 150, and a layout generation module 170. It may further have a conversion module 160. However, the storage module 150 does not need to be built into the information processing device 100, but only needs to be accessible from the information processing device 100 (call module 140). In other words, any information processing device 100 that can access the storage module 150 will suffice. As this form, for example, the storage module 150 may be located inside the information processing device 100, or the information processing device 100 and the storage module 150 may be in separate housings but connected to be accessible. As for the connection, if the storage module 150 is a device that can read a storage medium such as a USB memory or a CD, the connection with the device is via USB or the like, and if the storage module 150 is a server or the like, the connection is via a communication line such as the Internet. Furthermore, conversion module 160 is not an essential component of information processing device 100, and if conversion module 160 is not present, layout copy generation module 170 generates layout copies using the font generated by font generation module 130 and the font called by call module 140. If conversion module 160 is present, layout copy generation module 170 generates layout copies using the font generated by font generation module 130 and the font converted by conversion module 160.

[0035] The "'handwritten-style' font generated by the font generation module 130" is not the actual character written by a person by hand, nor is it a conventional typeface (a font with the same shape every time) that is printed, but rather a character that is close to the cursive style written by a person by hand. Specifically, it is a character that has a (moderate) fluctuation in the shape of the character (generally called "breaking"). Furthermore, a (moderate) fluctuation may be added between characters. More specifically, it is a font generated by a machine learning model. A "handwritten font stored in storage module 150" is a character generated based on a character written by hand. It may be a faithful vectorization of a handwritten character, or it may be a font generated by vectorizing a character drawn with a digital pen on a tablet device. Although it is a handwritten font, if this font is printed as is, it will be printed with the same font shape every time, and there will be no variation between multiple identical characters. Note that "variation" here refers to the difference in the shape (font) of a character when the same character (with the same character code) is written multiple times.

[0036] The receiving module 110 is connected to the classification module 120 and receives the character code. The characters to be targeted may include alphanumeric characters, hiragana, katakana, kanji, etc. used in Japanese, as well as blank characters (including tabs), symbols, emojis, characters used in other languages, figures, etc.

[0037] The classification module 120 is connected to the reception module 110, the font generation module 130, and the call module 140, and classifies the character code received by the reception module 110 into a predetermined first character and a predetermined second character. Note that the first character and the second character are both predetermined characters. The "first character" is a character for which a font can be generated by the font generation module 130, and the "second character" is a character other than the first character. In the specific example given below, the "first character" is a character for which a font can be dynamically generated using a machine-learned model, and the "second character" is a character for which the model cannot be used and for which a font in the storage module 150 is used. Here, the model is a machine-learned artificial intelligence (e.g., a neural network) that operates in a "prediction phase" that uses the artificial intelligence to provide a correct answer. Classifying means determining whether the accepted character code corresponds to the first character. If the character code corresponds to the first character, it is designated as the first character, and if it does not correspond to the first character, it is designated as the second character. Alternatively, if the character code corresponds to the second character, it is designated as the second character, and if it does not correspond to the second character, it is designated as the first character. Furthermore, if the character code corresponds to the first character, it is designated as the first character, and if the character code corresponds to the second character, it is designated as the second character. "Classifying into a predetermined first character and a predetermined second character" can be realized, for example, by (1) determining whether the character corresponds to a predetermined first character, (2) determining whether the character corresponds to a predetermined second character, or (3) determining whether the character corresponds to a predetermined first character and whether the character corresponds to a predetermined second character. Furthermore, as will be described later, a determination based on the type of accepted character code may also be added. Specifically, if the accepted character code is ASCII, which represents half-width alphanumeric characters, the character may be classified as a first character, and if it is UTF-16, which represents full-width characters, the aforementioned determination (one of (1), (2), or (3)) may be made. Furthermore, for example, the content of the "predetermined first character and second character" may be a first character and a second character classified by frequency of appearance. Here, the first character is a character that appears frequently, and the second character is a character other than the first character (a character that appears less frequently than the first character). Here, the frequency of appearance may be based on some statistical data. The statistical data may be statistical data compiled by a public institution or may be statistical data compiled by a private institution. Furthermore, the subject of the statistics may be, for example, the frequency of characters used in letters, or may be the frequency of characters used in newspapers, novels, emails, or characters previously accepted by this embodiment. The first characters may be alphanumeric characters and frequently occurring characters (in this case, frequently occurring hiragana, katakana, and kanji).The first characters may be alphanumeric characters, hiragana, katakana, and frequently occurring characters (in this case, frequently occurring kanji).

[0038] The font generation module 130 is connected to the classification module 120 and the layout generation module 170, and dynamically generates a handwritten font for the first character. "Dynamically generating a font" means generating a font (a handwritten-style font) that has fluctuations each time it is generated. Therefore, even when generating multiple fonts of the same character, the fonts will have different forms. However, even in this case, it is desirable that the difference between the two fonts is only enough to make them appear to have been written by the same person. A specific example is generating a font using a model that has undergone machine learning using handwritten characters written by humans as training data. For machine learning, a recurrent neural network (RNN), a deep learning method described in Non-Patent Document 1, may be used.

[0039] The call module 140 is connected to the classification module 120, the storage module 150, the conversion module 160, and the layout generation module 170, and calls a handwritten font stored in the storage module 150 for the second character. Specifically, the font corresponding to the character code of the second character is read from the storage module 150. The storage module 150 is connected to the call module 140 and stores handwritten fonts corresponding to character codes. The fonts stored in the storage module 150 (hereinafter also referred to as fixed fonts) are handwritten character fonts created in advance. As mentioned above, although they are handwritten fonts, the font data is fixed (constant). As will be described later, the conversion module 160 can dynamically convert these fixed fonts to make them look even more handwritten.

[0040] The layout copy generation module 170 is connected to the font generation module 130, the call module 140, the conversion module 160, and the writing device 180, and generates a layout copy using the font generated by the font generation module 130 and the font called by the call module 140. Therefore, at least in the font generated by the font generation module 130, there is fluctuation in the characters each time they are written, and since the font called by the call module 140 is a handwritten font, the overall impression is that the characters have fluctuation each time they are written. Also, as mentioned above, if the first character is a character that appears frequently, characters that have fluctuation each time they are written will occupy the majority of the block copy (written document), further creating a handwritten feel.

[0041] In addition, the font generation module 130 has a font generation module for each character type (font generation (A) module 132, font generation (B) module 134, font generation (C) module 136, etc.), and may dynamically generate a font using a font generation module corresponding to the character type of the first character. The character types may be, for example, alphanumeric characters, hiragana and katakana characters, and kanji characters. In this case, the "font generation modules for each character type" may be font generation (A) module 132, which is machine-learned using handwritten alphanumeric characters, font generation (B) module 134, which is machine-learned using handwritten hiragana and katakana characters, and font generation (C) module 136, which is machine-learned using handwritten kanji characters. Note that the three types of alphanumeric characters, hiragana and katakana characters, and kanji characters are examples, and other types may be used, for example, two types of alphanumeric characters and other characters (for example, hiragana, katakana, and kanji characters), or four types of alphanumeric characters, hiragana, katakana, and kanji characters, or may be divided into types other than these. For machine learning, for example, the RNN described above may be used.

[0042] In addition, the layout generation module 170 may assign information indicating the order of characters to be written to the font generated by the font generation module 130 and the font called by the call module 140 according to the order in which the character codes appear in the text received by the reception module 110. Specifically, the "information indicating the order of characters to be written" is generally a number, but alphabets or the like may also be used. By doing this, the character code and font do not need to be one-to-one correspondence, so even if you write the same font multiple times, you can write different font forms. Specifically, you can create a file for each font and use a sequential number in the file name, which is "information indicating the order in which the characters are written." For example, you can create files named 100.svg, 101.svg, etc.

[0043] The conversion module 160 is connected to the call module 140 and the layout generation module 170, and performs affine or homography conversion on the font called up by the call module 140. Since individual character variations are large in external shape, although the basic form is maintained, it is desirable to perform affine or homography conversion. In this case, the layout copy generation module 170 generates a layout copy using the font generated by the font generation module 130 and the font converted by the conversion module 160. The conversion module 160 makes the handwritten font stored in the storage module 150 resemble a handwritten font with fluctuations in the shape of the characters. A homography transformation is also called a projective transformation.

[0044] Furthermore, the coefficients in the affine transformation or homography transformation in the transformation module 160 may be changed by a random function. For example, "changing coefficients using a random function" can be done by changing the coordinates of four points around the periphery of a transformation quadrilateral used in an affine transformation or homography transformation so that they are scattered within a predetermined range using a random function. The "predetermined range" can be set to, for example, an amount determined from the appearance of a handwritten design (e.g., 5%). This "predetermined range" can create a sense of fluctuation in the shape while retaining the basic characteristics. 1 / f noise can also be used as the random function. Furthermore, as the writing distance (or writing time, number of characters written, etc.) increases, the "predetermined range" may be increased accordingly (to spread the values ​​more widely). This is to reproduce the fact that when many characters are written by hand, the character shapes often become distorted. Furthermore, towards the end of the composition (for example, when the remaining writing distance (or writing time, number of characters written, etc.) becomes smaller than the threshold), the "predetermined range" may be decreased (to prevent the values ​​from spreading more widely). This is to reproduce the fact that when writing by hand, the characters often become more carefully written near the end.

[0045] The coefficients may be changed to values ​​that are fluctuated by a random function either every time a font is generated or when the same character code appears in the layout. In other words, the coefficients can be converted each time a font is generated, or a determination can be made as to whether the same character code has previously existed within a single composition, and if so, the coefficients can be changed. In the former case, the coefficients will be different for each font. In the latter case, the same coefficients as the previous one will be used until the same character code appears. By changing the coefficients in this way, it is possible to mix character designs that are automatically generated by the font generation module 130 with character designs that are generated by subjecting a fixed font to affine or homography transformation, so that fluctuations occur in all characters each time they are written.

[0046] Furthermore, when writing vertically, the font of the alphanumeric character code may be rotated 90 degrees clockwise. This rotation process may be performed by the font generation module 130 or the layout generation module 170. For example, whether the text is written vertically or horizontally (or horizontally or vertically) can be specified by the user when inputting characters, and the receiving module 110 can receive and determine the specification.

[0047] The fonts included in the composition generated by the composition generation module 170 are written in vector format. This is a format called a vector font (also called a scalable font or stroke font). SVG (Scalable Vector Graphics) may be used as the composition data format. SVG is a type of image format that conforms to XML (a subset of XML) and is vector format data. To enable drawing with the writing device 180, it is necessary to express the line segment from its start point to its end point with a single smooth curve. To achieve this, a Bezier curve is used in XML. Note that a stroke font of a Bezier curve can be generated using the technology described in Patent Document 1.

[0048] The writing device 180 is connected to the layout generation module 170 of the information processing device 100, and has an output module 185 that writes characters according to vectors that indicate fonts in the layout generated by the layout generation module 170. The output module 185 is only required to be able to write characters on a medium (typically paper such as letter paper or postcards, but not limited to paper, and may also be cloth such as T-shirts, wood, etc.). Therefore, a printer (including a multifunction device) may be used, but a pen plotter (also called an XY plotter) that can use a pen is preferable to make the text appear handwritten. This output module 185 may also be called a robot, as it writes on behalf of a human. Examples of pens include fountain pens, ballpoint pens, writing brushes, felt-tip pens, and pencils. Although not limited thereto, a fountain pen is particularly preferable. When using a fountain pen, the fountain pen is fixed at an angle in the pen holder of the pen plotter. This vector includes at least the X and Y coordinates of the start and end points of the line segment (curve). It may also include the X and Y coordinates of the control points for creating a Bezier curve. Furthermore, if the output module 185 supports speed and pressure, it may also include writing speed and pressure.

[0049] Note that a space character (a space between words in the case of English) may be the first character or the second character. If it is the first character, if machine learning is performed including the space character, the font generation module 130 can impart fluctuation to the space character. If it is the second character, the conversion module 160 can impart fluctuation to the space character. Note that imparting fluctuation to a space character means changing the length of the space character (width in horizontal writing, height in vertical writing) each time (each time it appears). Furthermore, as will be described later, blank characters are treated as the first character, but processing may be further assigned based on the character code of the blank character. That is, if the blank character is a half-width alphanumeric character code, the font generation (A) module 132 processes it as the first character, and if the blank character is a full-width character code, the font generation (B) module 134 (which treats the blank character as one of the frequently occurring hiragana or katakana characters) or the font generation (C) module 136 (which treats the blank character as one of the frequently occurring kanji characters) processes it as the first character.

[0050] FIG. 2 is an explanatory diagram showing a schematic configuration example using this embodiment. As shown in the example of Fig. 2, this embodiment is a "system for robotic writing." That is, by processing the input subsystem 210, rendering subsystem 220, and recording robot subsystem 230 in that order, handwritten-style characters can be generated and used. The input subsystem 210 accepts input of character codes from a user who wants to generate a document that looks like handwritten characters. The rendering subsystem 220 generates a layout copy from the character code received by the input subsystem 210. This corresponds to the information processing device 100 described above. The recording robot subsystem 230 creates (prints) a document by writing handwritten-style characters according to the layout created by the rendering subsystem 220. This corresponds to the writing device 180 and output module 185 described above.

[0051] FIG. 3 is an explanatory diagram showing a more specific example of a configuration using this embodiment. This example system includes a user interface module 310, a typesetting control module 320, a font generation controller 330, an AI dynamic font generation engine 335, a calling module 340, a storage module 150, a font buffer 365, a rendering engine 370, a robot drive driver 380, and a pen plotter 390. The user interface module 310 includes the functions of the reception module 110, and through user operations, commands such as inputting character codes, displaying a preview of the layout (which is an on-screen display of the output image of the layout file 375 for confirmation), and having the robot driver 380 write.

[0052] The typesetting control module 320 includes the functions of the classification module 120, and controls the font generation controller 330 to generate a handwritten-style font according to the character code, the call module 340 to call the handwritten font in the fixed font storage module 350, and performs layout and generates the block copy file 375. The font generation controller 330 includes the functionality of the font generation module 130, and causes the AI ​​dynamic font generation engine 335 to generate handwritten fonts and store the handwritten fonts in the font buffer 365. The AI ​​dynamic font generation engine 335 includes functions such as the font generation (A) module 132, and generates handwritten-style fonts under the control of the font generation controller 330. Note that in the example of Fig. 3, there is one AI dynamic font generation engine 335, but as mentioned above, it is desirable to have multiple AI dynamic font generation engines 335 for different character types.

[0053] The call module 340 includes the functionality of the call module 140 and calls a handwritten font from the fixed font storage module 350. The call module 340 stores the handwritten font in the font buffer 365 or passes the handwritten font to the conversion module 360. The fixed font storage module 350 includes the functionality of the storage module 150 and stores handwritten fonts. The transformation module 360 ​​includes the functionality of the transformation module 160 and performs affine or homography transformations on handwritten fonts.

[0054] The font buffer 365 receives and stores fonts from the calling module 340 or the conversion module 360 ​​, and fonts from the font generation controller 330 . The rendering engine 370 generates an image from the font in the font buffer 365 and displays the image (preview image) on the user interface module 310. If the user checks the preview image and indicates approval based on the preview image, the rendering engine 370 passes a layout file 375 to the robot driver 380. The layout file 375 is, for example, an SVG file. The robot driver 380 causes the pen plotter 390 to draw fonts in accordance with the layout file 375. The pen plotter 390 is capable of recording and outputting from the layout file 375 which is typeset on one page.

[0055] FIG. 4 is an explanatory diagram showing an example of the configuration of a system using this embodiment. FIG. 4(a) shows an example of a stand-alone system configuration. The information processing device 100 is built into a personal computer 400 (a personal computer is an abbreviation for personal computer) (the program for the information processing device 100 is installed). The personal computer 400 and the pen plotter 390 are connected by, for example, a USB. The user operates the personal computer 400 to write a document on the pen plotter 390 .

[0056] FIG. 4(b) shows an example of a network-type system configuration. The layout creation server 410, personal computer 420A, and personal computer 420B are connected via a communication line 495. The communication line 495 may be wireless, wired, or a combination of these, and may be, for example, the Internet or an intranet as a communication infrastructure. The functions of the information processing device 100 may also be realized as a cloud service. The personal computer 420A and the pen plotter 390A are connected, and the layout creation server 410 and the pen plotter 390C are connected. The underprint creation server 410 incorporates the information processing apparatus 100 (installs the program of the information processing apparatus 100). The personal computer 400 and the pen plotter 390C are connected, for example, by USB. Note that a plurality of pen plotters 390 may be connected.

[0057] The user of the personal computer 420A connects to the underprint creation server 410 using a web browser or the like, inputs a character code, etc., and requests the underprint creation server 410 to perform processing. Then, the user receives the underprint generated by the underprint generation module 170 from the underprint creation server 410 and causes the pen plotter 390A to write a document. The user of the personal computer 420B connects to the underprint creation server 410 using a web browser or the like, inputs a character code, etc., and requests the underprint creation server 410 to perform processing. Then, the underprint creation server 410 causes the pen plotter 390C to write and mails the written document to the user of the personal computer 420B. Also, if it is a service that directly mails a postcard to the destination, the postcard written by the pen plotter 390C may be directly mailed to the destination. Note that although the personal computers 400 and 420 are used, not only laptop computers but also tablet terminals, smartphones, etc. may be used.

[0058] FIG. 5 is an explanatory diagram showing an example written according to the present embodiment. The writing example 500 in FIG. 5 shows an output example according to the present embodiment when the first character and the second character are mixed. That is, alphanumeric characters ( "A4", "Freehand") as the first character and characters with a high appearance frequency (hiragana, Chinese character "書") and a character with a low appearance frequency (Chinese character "絵") as the second character are written. Note that there are two line breaks in the characters input by the user (after "PADに" and after "絵が"), and a blank character (or tab) is inserted before the third line.

[0059] FIG. 6 is a flowchart showing a processing example according to the present embodiment. The font generation (A) module 132 is a model trained by machine learning using handwritten alphanumeric characters, the font generation (B) module 134 is a model trained by machine learning using handwritten hiragana and katakana, and the font generation (C) module 136 is a model trained by machine learning using handwritten kanji characters that appear frequently.

[0060] In step S602, the reception module 110 receives a character code in response to a user operation. In step S604, the classification module 120 determines whether the character code is a half-width alphanumeric character. If it is a half-width alphanumeric character (Y), the process proceeds to step S606. If it is not a half-width alphanumeric character (N, if it is a full-width character), the process proceeds to step S608.

[0061] In step S606, the font generation module 130 generates a handwritten font using model A (font generation (A) module 132). Note that the font is generated for each character string (word) of half-width alphanumeric characters. In step S608, classification module 120 determines whether the character code is a character stored in a table that stores characters with high occurrence frequency. If the character code is in the table (Y), proceed to step S610; otherwise (N, if the character is a character with low occurrence frequency), proceed to step S616. This table stores character codes for which fonts can be generated by the font generation (B) module 134 and font generation (C) module 136. The process in step S608 may be configured to determine whether the character code is a fixed font stored in the storage module 150. however In this case, "Y" and "N" in the flowchart are reversed. In other words, if the character code is a fixed font (if the character appears infrequently), proceed to step S616; otherwise (if the character appears frequently), proceed to step S610.

[0062] In step S610, classification module 120 determines whether the character code is "Hiragana or Katakana." If it is "Hiragana or Katakana" (Y), proceed to step S612; otherwise (N, if it is Kanji), proceed to step S614. In step S612, the font generation module 130 generates a hand-drawn style font using model B (font generation (B) module 134). In step S614, the font generation module 130 generates a hand-drawn style font using model C (font generation (C) module 136).

[0063] In step S616, the call module 140 calls the handwritten font corresponding to the character code from the storage module 150 (fixed font storage module 350). In step S618, the transformation module 160 modifies the coefficients of the affine or homography transformation. In step S620, the transformation module 160 performs an affine transformation or a homography transformation on the font called in step S616 using the coefficients changed in step S618 to generate a font.

[0064] In step S622, the layout creation module 170 assigns a number in the writing order to the file name, and stores the vector font (the font generated in steps S606, S612, S614, and S620) in that file. These processes (steps S602 to S622) are repeated the number of times equal to the number of received character codes. In step S624, a layout copy is generated using the file generated in step S622.

[0065] FIG. 7 is an explanatory diagram showing an example of processing according to this embodiment. The reception module 110 displays a print service screen 700 on the user's personal computer 400 or 420. The print service screen 700 displays a page type instruction field 710, a text input field 720, a file format instruction field 730, and a file creation button 740. The page type specification field 710 is a field for specifying the paper size, paper orientation (portrait or landscape), and whether the writing style is vertical or horizontal. The text input field 720 is a field for inputting text, that is, a character code accepted by the accepting module 110 is entered. The file format specification field 730 is a field for specifying the file format of the layout. Examples of file formats include SVG, PDF, etc. It also specifies whether a preview should be displayed. The file generation button 740 is a button that instructs the generation of a layout copy. When selected by the user, the layout copy is generated by processing the flowchart shown in the example of Figure 6, targeting the character code entered in the text input field 720. It should be noted that the print service screen 700 may have an area for displaying a preview.

[0066] The processing of the classification module 120, font generation module 130, call module 140, and layout generation module 170 will now be described in detail. The classification module 120 classifies half-width alphanumeric characters as a "word-by-word" character code string, and full-width characters as a character-by-character character code (see step S604 in the example of FIG. 6). As a specific example, half-width alphanumeric characters are ASCII, and full-width characters are UTF-16. If the input string is half-width alphanumeric characters, the classification module 120 transfers the processing to the font generation (A) module 132 for half-width alphanumeric characters, which generates a character image. If full-width characters (JIS-I, JIS-II level) are received, the classification module 120 transfers the data for each character to the font generation (B) module 134, font generation (C) module 136, or call module 140, which generates an SVG file for one character. The generated image files are given unique names such as 100.svg, 110.svg, etc. in the order of image generation, and are stored in memory (font buffer 365).

[0067] FIG. 8 is an explanatory diagram showing an example of processing according to this embodiment. If the target character code is a half-width alphanumeric character (one type of first character), the classification module 120 causes the alphanumeric character generation engine 810 (corresponding to the font generation (A) module 132) to process it. Furthermore, if the character code is full-width, the classification module 120 causes the first character (a character that appears frequently) to be processed by the handwritten-style hiragana / kanji generation engine 820 (corresponding to the font generation (B) module 134 or the font generation (C) module 136), and causes the second character (a character that appears less frequently) to be processed by the fixed font extraction module 830 (corresponding to the call module 140, or the call module 140 and the conversion module 160). Then, the layout generation module 170 generates a file called “number.svg” for the outputs of the alphanumeric character generation engine 810, the handwritten-style hiragana and kanji generation engine 820, and the fixed font extraction module 830.

[0068] FIG. 9 is an explanatory diagram showing an example of processing according to this embodiment. This shows an example of how character codes are allocated when "This section contains the message 'A Happy birthday to you.'" is entered in the text input field 720. The following line numbers are those in the processing example 900. The first line indicates that the character code for "this" is a Chinese character, which appears infrequently, and therefore the fixed font extraction module 830 is to process it. The second line indicates that the character code "section" is a kanji character and is a character that appears frequently, so that the handwritten-style hiragana and kanji generation engine 820 is to process it. The third line indicates that the character code for "wa" is a kanji character, which appears frequently, and therefore the handwritten-style hiragana and kanji generation engine 820 is to process it. The fourth line indicates that the character code "A" is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The fifth line indicates that the character code " " (half-width space character) is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The sixth line indicates that the character code string "happy" is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The seventh line indicates that the character code " " (half-width space character) is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The eighth line indicates that the character code string "birthday" is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The ninth line indicates that the character code " " (half-width space character) is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The tenth line indicates that the character code string "to" is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. Line 11 indicates that the character code " " (half-width space character) is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. The twelfth line indicates that the character code string "you." is a half-width alphanumeric character, and therefore the alphanumeric character generation engine 810 is to process it. Line 13 indicates that the character code " " (full-width space character) is a kanji (or hiragana) character that appears frequently, and therefore the handwritten hiragana / kanji generation engine 820 is to process it. The same process continues below. Naturally, if the character is full-width (i.e., not half-width) and has a low frequency of occurrence (the second character), it will be processed by fixed font extraction module 830. In the example of FIG. 9, only one character is processed by fixed font extraction module 830, while the others are processed by alphanumeric character generation engine 810 or handwritten-style hiragana / kanji generation engine 820, so the overall impression is perceived as handwritten. In other words, the font generated by fixed font extraction module 830 is originally a handwritten font and has a low frequency of occurrence in the generated document (the probability of the same character appearing multiple times in a single document is even lower), reducing the mechanical impression. Furthermore, as mentioned above, applying affine transformation or the like to the font generated by fixed font extraction module 830 can further reduce the mechanical impression.

[0069] Although the above-described embodiment mainly shows an example of output using a pen plotter, output using a printer is also possible. For example, the main text of a document may be printed using a typeface, while portions such as the addressee and signature may be printed using a font generated in this embodiment. Because it is a printer, the handwritten feel is less pronounced than with a pen plotter, but it can be made to feel more familiar than an all-typeface font. Specifically, character codes within predetermined areas (areas for the addressee, signature, etc.) are processed using this embodiment, and character codes within other areas (the main text area) are printed using a conventional typeface font.

[0070] The hardware configuration of the computer on which the program according to this embodiment is executed is a general computer, specifically a personal computer, a computer that can be a server, etc., as shown in Fig. 10. That is, as a specific example, a processor, CPU 1001, is used as a processing unit (arithmetic unit), and RAM 1002, ROM 1003, and HD 1004 are used as storage devices (corresponding to storage module 150). HD 1004 may be, for example, a hard disk or SSD (Solid State Drive). The computer system is comprised of a CPU 1001 that executes programs such as a reception module 110, a classification module 120, a font generation module 130, a call module 140, a conversion module 160, a layout generation module 170, and a writing device 180, a RAM 1002 that stores the programs and data, a ROM 1003 that stores programs for starting up the computer, a HD 1004 that is an auxiliary storage device (which may be a flash memory or the like) that stores the processing results of the modules, a reception device 1006 that receives data based on user operations (including movements, voice, and gaze) on a keyboard, mouse, touch screen, microphone, camera (including a video camera, a gaze detection camera, etc.), etc., an output device 1005 such as an LCD display, an organic EL display, a projector, speakers, or a pen plotter, a communication line interface 1007 such as a network interface card for connecting to a communication network, and a bus 1008 that connects them to exchange data. A plurality of these computers may be connected to one another via a network.

[0071] Of the above-mentioned embodiments, those that are implemented using computer programs are realized by loading the computer program, which is software, into a system with this hardware configuration, and the software and hardware resources work together to realize the above-mentioned embodiments. 10 shows only one example of the hardware configuration, and the present embodiment is not limited to the configuration shown in FIG. 10, and any configuration capable of executing the modules described in the present embodiment may be used. For example, some modules may be configured with dedicated hardware (e.g., an Application Specific Integrated Circuit (ASIC) or the like), some modules may be in an external system and connected via a communication line, or a plurality of the systems shown in FIG. 10 may be connected to each other via a communication line and operate in cooperation with each other. In particular, the systems may be incorporated into personal computers, portable information communication devices (including mobile phones, smartphones, mobile devices, wearable computers, etc.), home information appliances, robots, etc.

[0072] The program described above may be provided by being stored on a recording medium, or may be provided via communication means. In such cases, the program described above may be considered as an invention of a "computer-readable recording medium on which a program is recorded." "Computer-readable recording medium on which a program is recorded" means a computer-readable recording medium on which a program is recorded, which is used for installing, executing, distributing, etc. a program. Recording media include, for example, digital versatile discs (DVDs) such as DVD-R, DVD-RW, DVD-RAM, etc., which are standards established by the DVD Forum, and DVD+R, DVD+RW, etc., which are standards established for DVD+RW; compact discs (CDs) such as read-only memory (CD-ROM), CD recordable (CD-R), CD rewritable (CD-RW), Blu-ray (registered trademark) Disc, magneto-optical disk (MO), flexible disk (FD), magnetic tape, hard disk, read-only memory (ROM), electrically erasable and rewritable read-only memory (EEPROM), flash memory, random access memory (RAM), SD (short for Secure Digital) memory cards, etc. The program, in whole or in part, may be recorded on the recording medium and stored, distributed, etc. It may also be transmitted by communication using a transmission medium such as a wired network used in a local area network (LAN), metropolitan area network (MAN), wide area network (WAN), the Internet, an intranet, an extranet, etc., a wireless communication network, or a combination thereof, or may be carried on a carrier wave. Furthermore, the program may be a part or all of another program, or may be recorded on a recording medium together with a separate program. It may also be split and recorded on multiple recording media. It may also be recorded in any format, such as compressed or encrypted, as long as it is restorable. [Explanation of symbols]

[0073] 100...Information processing device 110...Reception module 120...Classification module 130...Font generation module 132...Font generation (A) module 134...Font generation (B) module 136...Font generation (C) module 140...Call module 150...Memory module 160...Conversion module 170...Plate generation module 180...Writing device 185...Output module 210...Input subsystem 220...Rendering subsystem 230...Recording robot subsystem 310...User Interface Module 320...Typesetting control module 330...Font generation controller 335…AI dynamic font generation engine 340...Call module 350...Fixed font memory module 360...Conversion module 365...Font buffer 370...Rendering engine 375...Plate file 380...Robot driver 390...pen plotter

Claims

1. Accepting means for accepting a character code; a classification means for performing one of the following processes for the character code received by the receiving means: (1) if the character code corresponds to a predetermined first character code, classifying the character code as a first character, and if it does not correspond to the first character code, classifying the character code as a second character; (2) if it corresponds to a predetermined second character code, classifying the character code as a second character, and if it does not correspond to the second character code, classifying the character code as a first character; or (3) if it corresponds to the predetermined first character code, classifying the character code as a first character, and if it corresponds to the predetermined second character code, classifying the character code as a second character; a font generating means for dynamically generating a handwritten font for the first character; a calling means for calling a handwritten font stored in a storage means for the second character; a block copy generating means for generating a block copy using the font generated by said font generating means and the font called by said calling means; An information processing device having the above.

2. the font generation means has a font generation means for each character type, and dynamically generates a font using the font generation means corresponding to the character type of the first character; The information processing device according to claim 1 .

3. The block copy generating means assigns information indicating the order of characters to be written to the font generated by the font generating means and the font called by the calling means in accordance with the order of appearance of character codes in the text received by the receiving means. The information processing device according to claim 1 .

4. a conversion means for performing an affine transformation or a homography transformation on the font called by the calling means; and the block copy generating means generates a block copy using the font generated by the font generating means and the font converted by the converting means; The information processing device according to claim 1 .

5. changing coefficients in the affine transformation or the homography transformation by the transformation means using a random function; The information processing device according to claim 4 .

6. The timing for changing the coefficient to a value fluctuated by a random function is either every time a font is generated or when the same character code appears in the layout. The information processing device according to claim 5 .

7. When writing vertically, the font of the alphanumeric character code is rotated 90 degrees clockwise. The information processing device according to claim 1 .

8. A writing system having an information processing device and a writing device according to any one of claims 1 to 7, the font is in vector format; the writing device has an output means for writing characters in accordance with vectors representing the font in the layout created by the layout creation means; Writing system.

9. Computer, Accepting means for accepting a character code; a classification means for performing one of the following processes for the character code received by the receiving means: (1) if the character code corresponds to a predetermined first character code, classifying the character code as a first character, and if it does not correspond to the first character code, classifying the character code as a second character; (2) if it corresponds to a predetermined second character code, classifying the character code as a second character, and if it does not correspond to the second character code, classifying the character code as a first character; or (3) if it corresponds to the predetermined first character code, classifying the character code as a first character, and if it corresponds to the predetermined second character code, classifying the character code as a second character; a font generating means for dynamically generating a handwritten font for the first character; a calling means for calling a handwritten font stored in a storage means for the second character; a block copy generating means for generating a block copy using the font generated by said font generating means and the font called by said calling means; An information processing program that functions as a

10. An information processing method performed by an information processing device that can access a storage means, The information processing device performs the following steps: an acceptance step for accepting a character code; a classification step for performing one of the following processes for the character code received in the receiving step: (1) if the character code corresponds to a predetermined first character code, classifying the character code as a first character, and if it does not correspond to the first character code, classifying the character code as a second character; (2) if it corresponds to a predetermined second character code, classifying the character code as a second character, and if it does not correspond to the second character code, classifying the character code as a first character; or (3) if it corresponds to the predetermined first character code, classifying the character code as a first character, and if it corresponds to the predetermined second character code, classifying the character code as a second character; a font generation step of dynamically generating a handwritten font for the first character; a recall step of recalling a handwritten font stored in the storage means for the second character; a block copy generating step for generating a block copy using the font generated in the font generating step and the font called in the calling step; An information processing method comprising:

Images