JP2021106397A - Server and data allocation method - Google Patents

Abstract

To provide a server and a data allocation method (program) that enable easy creation of composite content data.SOLUTION: A server 1, a manager terminal 2, and a user terminal 3 are communicatively connected to each other through a network in a system. The server 1 includes: a second data generation unit 130 that receives input of first data including at least identified information and generates one or more pieces of second data including at least a part of the identified information; a second data allocation unit 142 that allocates the second data to a cut of third data including one or more cuts; an associating unit 150 that associates the second data and material content data on the basis of at least a part of the identified information included in the second data and extracted information extracted from the material content data; and a material content data allocation unit 144 that allocates the material content data to the cut on the basis of the association.SELECTED DRAWING: Figure 4

Description

The present invention relates to a server that generates composite content data and a data allocation method.

Conventionally, content data such as moving images has been created. For example, Patent Document 1 proposes a moving image processing device that efficiently searches for a desired scene image from moving images having a plurality of chapters.

Japanese Unexamined Patent Publication No. 2011-130007

Creating content data such as moving images takes a lot of time and effort, and in particular, when creating composite content data using multiple material content data such as text data, images, and sound data, the optimum combination is used. Since it is difficult to take this into consideration depending on the technical level of the user, it has been required to provide a device that can easily create composite content data.

Therefore, an object of the present invention is to provide a server and a program that enable easy creation of composite content data.

The main invention of the present invention for solving the above problems is a first method of accepting input of first data including at least the identified information and generating one or more second data including at least a part of the identified information. At least the second data generation unit, the second data allocation unit that allocates the second data to the cut of the third data including one or more cuts, and the identified information included in the second data. Based on a part and the extracted information extracted from the material content data, the association unit that associates the second data with the material content data, and the material content that allocates the material content data to the cut based on the association. It is a server characterized by including a data allocation unit.

According to the present invention, it is possible to provide a server and a data allocation method that enable easy creation of composite content data.

It is a block diagram of the system which concerns on embodiment. It is a block diagram of the server which concerns on embodiment. It is a block diagram of the management terminal and the user terminal which concerns on embodiment. It is a functional block diagram of the system which concerns on embodiment. It is a figure explaining the screen layout example which constitutes the 3rd data. It is a flowchart of the system which concerns on embodiment. It is explanatory drawing of the aspect of displaying a list of a plurality of cuts constituting the composite content data on the screen. This is the processing flow of the summary sentence creation function. It is a figure explaining the process of subjecting a document to morphological analysis and dividing it into tokens. It is a figure explaining the insertion of the sentence into each cut of the 3rd data.

The contents of the embodiments of the present invention will be described in a list. The server and the data allocation method according to the embodiment of the present invention have the following configurations.
[Item 1]
A second data generation unit that accepts input of first data including at least the identified information and generates one or more second data including at least a part of the identified information.
A second data allocation unit that allocates the second data to the cut of the third data including one or more cuts, and a second data allocation unit.
An association unit that associates the second data with the material content data based on at least a part of the identified information included in the second data and the extracted information extracted from the material content data.
Based on the association, the material content data allocation unit that allocates the material content data to the cut, and
A server characterized by being equipped with.
[Item 2]
The identified information is a vector representing a word form or an entire sentence defined by a language model.
The server according to item 1, characterized in that.
[Item 3]
The first data is at least one of text data including the identified information, image data or moving image data including the identified information, and audio data including the identified information.
One,
The server according to item 1 or 2, characterized in that.
[Item 4]
The second data analyzes the maximum number of characters in each cut of the third data and the modification relationship between the clauses of the first data, and divides the sentence so that a natural section as a sentence fits in each cut. It is
The server according to items 1 to 3, wherein the server is characterized by the above.
[Item 5]
The material content data is at least one of image data, video data, and sound data.
One,
The server according to items 1 to 4, wherein the server is characterized by the above.
[Item 6]
The associating unit compares at least a part of the identified information included in the second data with the extracted information extracted from the material content data, determines the degree of similarity to each other, and determines the degree of similarity to each other. The material content data suitable for the data and the second data are associated with each other.
The server according to items 1 to 5, wherein the server is characterized by the above.
[Item 7]
It is a trained model in which machine learning is performed using the training data, and is provided with a classifier that extracts the extracted information of the material content data.
The extracted information is a class label output by the classifier.
The server according to item 6, characterized in that.
[Item 8]
The server further has a function of summarizing text information contained in the first data.
The second data generation unit generates the second data from the summary.
The server according to any one of items 1 to 7, characterized in that.
[Item 9]
The material content data is acquired from the network and stored in the material content data storage unit.
The server according to any one of items 1 to 8, characterized in that.
[Item 10]
A step of accepting input of first data including at least the identified information and generating one or more second data including at least a part of the identified information.
A step of assigning the second data to the cut of the third data including one or more cuts,
A step of associating the second data with the material content data based on at least a part of the identified information included in the second data and the material content data.
Including a step of assigning the material content data to the cut based on the association.
A data allocation method characterized by that.

<Details of the embodiment>
Hereinafter, a system for creating composite content data according to an embodiment of the present invention (hereinafter referred to as “the present system”) and the like will be described. In the accompanying drawings, the same or similar elements are given the same or similar reference numerals and names, and duplicate description of the same or similar elements may be omitted in the description of each embodiment. In addition, the features shown in each embodiment can be applied to other embodiments as long as they do not contradict each other.

<Structure>
As shown in FIG. 1, the system according to the embodiment is configured to include a server 1, an administrator terminal 2, and a user terminal 3. The server 1, the administrator terminal 2, and the user terminal 3 are connected to each other so as to be able to communicate with each other via a network. The network may be a local network or may be connectable to an external network. In the example of FIG. 1, an example in which one server 1 is configured is described, but the server 1 is composed of a plurality of server devices.
It is also possible to realize. Further, the server 1 and the administrator terminal 2 may be shared.

<Server 1>
FIG. 2 is a diagram showing a hardware configuration of the server 1 shown in FIG. The illustrated configuration is an example, and may have other configurations. Further, the server 1 may be a general-purpose computer such as a workstation or a personal computer, or may be logically realized by cloud computing.

The server 1 has at least a processor 10, a memory 11, a storage 12, and a transmitter / receiver 1.
3. Input / output units 14 and the like are provided, and these are electrically connected to each other through a bus 15.

The processor 10 is an arithmetic unit that controls the operation of the entire server 1, controls the transmission and reception of data between each element, and performs information processing necessary for application execution and authentication processing. For example, the processor 10 is a CPU (Central Processing Unit).
) And GPU (Graphics Processing Unit), and execute each information processing by executing a program or the like for this system stored in the storage 12 and expanded in the memory 11. The processing capacity of the processor 10 may be sufficient to execute necessary information processing. Therefore, for example, the processor 10 may be composed of only a CPU, and is not limited to this.

The memory 11 is a DRAM (Dynamic Random Access Memory).
Main memory composed of volatile storage devices such as y), flash memory and HDD (Hard)
Includes auxiliary storage configured with a non-volatile storage device such as Disc Drive). The memory 11 may be used as a work area of the processor 10 or the like, and may store a BIOS (Basic Input / Output System) executed when the server 1 is started, various setting information, and the like.

The storage 12 stores various programs such as application programs. A database storing data used for each process may be built in the storage 12.

The transmission / reception unit 13 connects the server 1 to the network.

The input / output unit 14 is an information input device such as a keyboard and a mouse, and an output device such as a display.

The bus 15 is commonly connected to each of the above elements and transmits, for example, an address signal, a data signal, and various control signals.

<Administrator terminal 2, user terminal 3>
The administrator terminal 2 and the user terminal 3 shown in FIG. 3 also include a processor 20, a memory 21, a storage 22, a transmission / reception unit 23, an input / output unit 24, and the like, and these are electrically connected to each other through a bus 25. .. Since the functions of each element can be configured in the same manner as the server 1 described above, detailed description of each element will be omitted. The administrator uses the administrator terminal 2 to change the settings of the server 1 and manage the operation of the database, for example. The user can access the server 1 by the user terminal 3 and browse the created composite content data, for example.

<Function of server 1>
FIG. 4 is a block diagram illustrating the functions implemented in the server 1. In the present embodiment, the server 1 is the communication unit 110, the identified information analysis unit 120, and the second data generation unit 130.
, Composite content data generation unit 140, association unit 150, storage unit 160, classifier 170
It has. The composite content data generation unit 140 is the second data allocation unit 142,
The material content data allocation unit 144 is included. Further, the storage unit 160 includes a third data storage unit 162, a material content data storage unit 164, and a composite content data storage unit 166.
Includes various databases such as the interface information storage unit 168.

The communication unit 110 communicates with the management terminal 2 and the user terminal 3. The communication unit 110 also functions as a reception unit that receives at least the first data including the identified information from the user terminal 2. The first data is, for example, text data such as an article containing the identified information (for example, a press release or news), image data including the identified information (for example, a photograph or an illustration) or moving image data. , Voice data including identified information may be used. The text data referred to here is not limited to the text data at the time of being transmitted to the server 1, for example, the voice data transmitted to the server 1 is text data generated by a known voice recognition technique. You may. Further, the first data may be, for example, text data such as an article summarized by an existing automatic summarization technique such as an extractive summarization or a generative summarization (including identified information). In this case, the number of cuts included in the third data can be reduced, the data capacity of the entire composite content data can be reduced, and the content can be simplified.

Further, the voice data referred to here may be voice data extracted from moving image data by voice recognition or voice data generated from text data. In the former case, only audio data such as narration and dialogue is extracted from the temporary video such as a temporary image such as a rough sketch and a video based on the temporary video, and as will be described later, the composite content is combined with the material content data based on the audio data. Data may be generated. In the latter case, for example
Audio data may be created from text data with a story, and in the case of a fairy tale, for example, a picture-story show or a moving image based on the story read aloud and the material content data may be generated as composite content data.

When the second data generation unit 130 determines that it is not necessary to divide the first data (for example, the text data is a short sentence having a preset number of characters or less), the second data generation unit 130 is used for the second data. The data generation unit 130 directly generates the first data as the second data. On the other hand, for example, when it is determined that the first data needs to be divided (for example, the sentence is longer than the preset number of characters), the second data generation unit 130 uses the first data generation unit 130.
Data is divided and generated as second data including at least a part of the identified information of the first data. The first data division method by the second data generation unit 130 may use any known technique, for example, as long as the first data can be converted into text, as will be described later. , The maximum number of characters in each cut of the third data and the modification relationship between clauses may be analyzed, and the sentences may be separated so that a natural section as a sentence fits in each cut.

The identified information analysis unit 120 analyzes the above-mentioned second data and acquires the identified information. Here, the identified information may be any information as long as it can be analyzed by the identified information analysis unit 120. In one embodiment, the identified information can be in the word form defined by the language model. More specifically, one or more words with the word vector described below (eg, ""
It may be "Shibuya, Shinjuku, Roppongi" or "Shibuya, landmark, youth"). It should be noted that the word may include a word that is not normally used by itself, such as "n", depending on the language model. Further, instead of the above word format, it may be a document accompanied by a vector representing the entire sentence, or a feature vector extracted from an image or a moving image.

The composite content data generation unit 140 reads out the third data including one or more cuts from the third data storage unit 162, and the material content data stored in the material content data storage unit 164 and the above-mentioned second data are combined. The assigned third data is generated as composite content data and stored in the composite content data storage unit 166, and the user terminal 2
Display composite content data in. Note that FIG. 5 is an example of the screen layout of the cuts constituting the third data. The edited second data (for example, a delimited text sentence) is inserted into the second data field 31 in the figure, and the selected material content data is inserted into the material content data field 32. For each cut of the third data
The preset maximum number of characters (in the case of text data), screen layout, and playback time (in the case of moving images) are specified. Further, the composite content data does not necessarily have to be stored in the composite content data storage unit 166, and may be stored at an appropriate timing. Further, the third data to which only the second data is assigned may be displayed on the user terminal 3 as progress information of the composite content data.

The second data allocation unit 142 assigns numbers to a plurality of cuts, such as cut 1, cut 2, and cut 3, and assigns the second data in this numerical order. The allocation method is not limited to this, and for example, the second data allocation unit 1
42 may increase or decrease the number of cuts included in the third data from a preset number according to the number of the second data, and allocate the second data.

The association unit 150 includes at least a part of the identified information included in the second data described above and the association unit 150.
For example, the material content data suitable for the second data (for example, by comparing with the extracted information extracted from the material content data (for example, the class label extracted by the classifier) and determining the similarity with each other, for example. (For example, those having a high degree of similarity) and the second data are associated with each other. As a more specific example, for example, the identified information included in the second data is ".
Material content data A that represents "teacher" and whose extracted information is "face" (for example, an image of a woman)
When the material content data B (for example, an image of Mt. Fuji) that is "mountain" is prepared, the relationship between the word vector obtained from "teacher" and the word vector obtained from "face" is obtained from "teacher". The second data is associated with the material content data A because it is more similar than the association between the word vector to be obtained and the word vector obtained from the "mountain". The extraction information of the material content data may be extracted in advance by the user and stored in the material content data storage unit 164, or may be extracted by the classifier 170 described later.
Further, in the determination of the similarity, a trained model in which a word vector is learned may be prepared, and the similarity of words may be determined by a method such as cosine similarity or Word Mover's Distance using the vector.

The material content data may be, for example, image data, moving image data, sound data (for example, music data, audio data, sound effects, etc.), but is not limited thereto. Further, the material content data may be stored in the material content data storage unit 164 by the user or the administrator, or the material content data is acquired from the network and stored in the material content data storage unit 164. It may be.

The material content data allocation unit 144 allocates suitable material content data to the cut to which the corresponding second data is assigned based on the above-mentioned association.

The interface information storage unit 168 stores various control information for display on the display unit (display or the like) of the management terminal 2 or the user terminal 2.

The classifier 170 is created as a trained model by acquiring training data from a training data storage unit (not shown) and performing machine learning. The creation of the classifier 170 is done on a regular basis (
For example, several times a year). As the training data for creating the classifier, data collected from the network or data owned by the user with a class label may be used, or a data set with a class label may be procured and used. .. Then, the classifier 170 is, for example, a trained model using a convolutional neural network, and when material content data is input, one or a plurality of extraction information (for example, a class label) is extracted. The classifier 170 extracts, for example, class labels (eg, seafood, grilled meat, people, furniture) representing objects related to material content data.

FIG. 6 is a diagram illustrating an example of a flow of creating composite content data.

First, the server 1 receives the first data including at least the identified information from the user terminal 2 via the communication unit 110 (step S101). In this example, the identified information may be, for example, one or more words, and the first data may be, for example, text data consisting of an article containing one or more words or a summary of the text data.

Next, the server 1 reads the third data including one or more cuts from the third data storage unit 162 by the composite content data generation unit 140 (step S102). In this example, the third data can be, for example, template data including a plurality of cuts.

Next, the server 1 analyzes the first data by the identified information analysis unit 120 to acquire the identified information, and the second data generation unit 130 includes at least a part of the identified information. The second data of the above is generated (step S103). The second data may be individual text data in which the maximum number of characters in each cut of the third data and the modification relationship between clauses are analyzed, and the sentences are divided so that a natural section as a sentence fits in each cut.

Next, the server 1 allocates the second data to the cut by the second data allocation unit (step S104). The progress may be confirmed by displaying the third data in this state on the user terminal 3.

Next, the server 1 is based on at least a part of the identified information included in the second data and the extracted information extracted from the material content data, and the association unit 150 causes the material content data of the material content data storage unit 164. And the second data are associated with each other (step S105), and the material content data is assigned to the cut by the material content data allocation unit 144 (step S106).

Then, the server 1 generates the second data and the third data to which the material content data is assigned as the composite content data, and the composite content data storage unit 16
It is stored in No. 6 and the composite content data is displayed on the user terminal 2 (step S107). As for the display of the composite content data, as illustrated in FIG. 7, a plurality of cuts constituting the composite content data can be displayed in a list on the screen. In each cut, information on the reproduction time (seconds) of each cut may be displayed together with the displayed material content data and the second data. The user can modify the content by clicking, for example, the second data field 31 or the corresponding button, and replace the material content data by clicking the material content data field 32 or the corresponding button. Can be done.

The flow of creating the above-mentioned composite content data is an example. For example, step S102 for reading the third data is executed when the second data or the material content data has been read before being assigned. It may have been done. Further, for example, the order of step S104 for allocating the second data, step S105 for associating, and step S106 for allocating the material content data is also executed in any order as long as there is no discrepancy with each other. May be done.

(Summary sentence creation function)
With reference to FIGS. 8 to 10, when the first data includes text data, the server 1 creates a summary sentence from the text data, and the second data from the first data including the summary sentence is second.
An example of creating the data of is described.

STEP91: Paragraph division / document division The server 1 divides the input text data document into paragraphs and divides the document in each paragraph into sentences. Further, a sentence (for example, 80 characters or more) that is too long to be displayed in one scene as a telop of a moving image is further divided into a plurality of sentences at a place satisfying a specific part of speech, notation, and the like.

STEP92: Document morphological analysis server 1 performs morphological analysis on each sentence and extracts a token which is the minimum unit of parsing. As shown in FIG. 9, each token is given a part of speech.

STEP93: Deletion of unnecessary words / unnecessary paragraphs The server 1 deletes sentences and paragraphs defined as invalid from the predetermined invalid sentence determination rule. For example, delete a line starting with a specific symbol such as "■" or "▼", a paragraph surrounded by a specific symbol, a URL, an email address, an address / telephone number, or the like.

STEP94: Deletion of stop words, etc. Server 1 deletes specific part of speech such as words (stop words) and particles that are not very important, such as "ni", "kara", "kore", and "san", from the token. ..

STEP95: Creating a token biggram A token biggram is obtained by connecting a plurality of tokens satisfying a specific condition (for example, a predefined part of speech condition). For example, "2014" (nouns, proper nouns, general) and "June" (June)
Connect nouns, proper nouns, and general) to make "June 2014", or connect "Verdi" (proper noun) and "sponsor" (common noun) to make "Verdi sponsor".

STEP96: Extraction of important sentences Extract the tokens and token viagrams that are characteristic words from the index that evaluates the importance of words such as TF-IDF based on the tokens and token bigrams, and segment the text from the above-mentioned word similarity judgment. And extract important sentences from each segment to make a summary.

STEP97: Application to template The abstract (important sentence) is parsed and separated into clauses and syntax trees. In the above-mentioned third data (for example, template data), the number of characters that can be inserted into each cut is defined, and the sentence is divided so that a natural section as a sentence fits in each template due to the modification relationship between clauses, and the third data. Apply to the data of. FIG. 10 shows an example of inserting a sentence into each cut of the third data.
The summary sentence creation function described above can support not only Japanese but also multiple languages including English.

According to this system of the embodiment described above, it is possible to easily create composite content data without having to prepare editing software, a server, an editor with specialized skills, and the like. For example, it is expected to be used in the following situations.
1) Video of product information sold at EC shops 2) Video distribution of press release information, CSR information, etc. 3) Video of manuals such as usage and operation flow 4) Creation of creatives that can be used as video advertisements

Although the preferred embodiment of the present invention has been described above, the technical scope of the present invention is not limited to the description of the above embodiment. Various changes / improvements can be made to the above-described embodiment, and those in which such changes / improvements have been made are also included in the technical scope of the present invention.

1 Server 2 Administrator terminal 3 User terminal

Claims (10)

A second data generation unit that accepts input of first data including at least the identified information and generates one or more second data including at least a part of the identified information.
A second data allocation unit that allocates the second data to the cut of the third data including one or more cuts, and a second data allocation unit.
An association unit that associates the second data with the material content data based on at least a part of the identified information included in the second data and the extracted information extracted from the material content data.
Based on the association, the material content data allocation unit that allocates the material content data to the cut, and
A server characterized by being equipped with. The identified information is a vector representing a word form or an entire sentence defined by a language model.
The server according to claim 1. The first data is text data including the identified information.
The server according to claim 1 or 2. The second data analyzes the maximum number of characters in each cut of the third data and the modification relationship between the clauses of the first data, and divides the sentence so that a natural section as a sentence fits in each cut. It is
The server according to claim 1 to 3. The material content data is at least one of image data, video data, and sound data.
One,
The server according to claim 1 to 4. The associating unit compares at least a part of the identified information included in the second data with the extracted information extracted from the material content data, determines the degree of similarity to each other, and determines the degree of similarity to each other. The material content data suitable for the data and the second data are associated with each other.
The server according to claim 1 to 5. It is a trained model in which machine learning is performed using the training data, and is provided with a classifier that extracts the extracted information of the material content data.
The extracted information is a class label output by the classifier.
The server according to claim 6. The server further has a function of summarizing text information contained in the first data.
The second data generation unit generates the second data from the summary.
The server according to any one of claims 1 to 7. The material content data is acquired from the network and stored in the material content data storage unit.
The server according to any one of claims 1 to 8. A step of accepting input of first data including at least the identified information and generating one or more second data including at least a part of the identified information.
A step of assigning the second data to the cut of the third data including one or more cuts,
A step of associating the second data with the material content data based on at least a part of the identified information included in the second data and the material content data.
Including a step of assigning the material content data to the cut based on the association.
A data allocation method characterized by that.

Images