JP4652357B2 - Media editing server device, media editing system, and media editing program - Google Patents

Description

  The present invention relates to a technology of a media editing system that is a moving image or a sound, and more particularly to a technology of generating an editing graph for editing media according to the described order.

Conventionally, video editing has been performed using a video editing system in which expensive editing software is installed in a dedicated equipment or a high-spec PC (personal computer). Therefore, when shooting with a camera installed on a mobile phone or the like, it was necessary to move the shot video to the video editing system and edit the video shot on the video editing system. I couldn't edit it right away even when shooting with the camera.
Therefore, a method of editing using an online application instead of an application on a terminal has been proposed.

At that time, since the general user is not necessarily detailed in the editing technique, the editing operation is not performed by a complicated program, but the user downloads a scenario template described in XML (Extensible Markup Language) or the like. In the downloaded scenario template, shooting time and camera movement are determined and recorded in advance.
The user performs shooting according to the shooting time and camera movement recorded in the downloaded scenario template. Thereafter, the user transmits the shot video together with the scenario template to the video editing server. A technology of a personal video production system is known in which a video editing server edits a video received based on a scenario template received from a user, so that the user can easily generate a video (for example, non-patented). Reference 1).
D-9-9 "Personal video production system using vlog template", NTT Communication Solutions Laboratories NTT Cyber Solutions Laboratories NTT Telecom Corporation, Hiroshi Konishi, Masashi Morimoto

However, the personal video production system shown in Non-Patent Document 1 has a problem that it is impossible to generate a scenario template in which the order for editing video is described.
For this reason, the personal video production system shown in Non-Patent Document 1 has a problem that the user can edit video only with a scenario template generated in advance.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a media editing server device and media capable of generating a scenario template in which an order for editing media that is moving images or audio is described. To provide an editing system and a media editing program.

  The present invention has been made to solve the above-described problems, and the invention according to claim 1 relates to an API that processes media that is a moving image or audio, and a basic API that has different API parameter values. , A basic API correspondence storage unit stored for each API, a transition probability storage unit storing transition probabilities between the basic APIs, and a plurality of APIs being input in response to the input A basic API selection unit that reads and outputs a basic API corresponding to a plurality of APIs from the basic API correspondence storage unit, and reads a transition probability between the basic APIs input from the basic API selection unit from the transition probability storage unit, A graph estimation unit that generates an edit graph to which the basic API is connected based on the read transition probability, and an edit graph generated by the graph estimation unit A media editing server apparatus characterized by having an execution unit for editing the media based.

  According to a second aspect of the present invention, the media editing server device includes an editing graph accumulating unit that stores an editing graph in which the order of the APIs and setting values of the parameters of the API are described, and the editing graph accumulating unit Based on the edit graph stored in the section, for each API, the API is classified into the plurality of basic APIs having statistically different parameter setting values, and the APIs are associated with the classified basic APIs. Calculating a transition probability between the basic API classified by the basic API classification unit and the basic API classified by the basic API classification unit based on the edit graph stored in the edit graph storage unit, The medium according to claim 1, further comprising: a basic API transition probability calculation unit that stores the calculated transition probability in the transition probability storage unit. As a current server device.

  According to a third aspect of the present invention, there is provided a classification information storage unit storing classification information for classifying setting values of a plurality of parameters used by the media editing server device to classify APIs into basic APIs. And having the basic API classifying unit accumulate classification information for statistically classifying the setting values of the plurality of parameters when the API classifies the API into the basic API. Calculated based on a graph, classifies the API into the basic API based on the calculated classification information, stores the calculated classification information in the classification information storage unit, and the basic API selection unit includes the API and In response to the input of the API parameter setting values, the classification information corresponding to the input API is read from the classification information storage unit, and the read Based on the class information and the set values of the input parameters, the set values of the plurality of input parameters from the basic API read out from the basic API corresponding storage unit corresponding to the plurality of input APIs. 3. The media editing server device according to claim 2, wherein a basic API corresponding to is selected and the selected basic API is output.

The invention according to claim 4 is a media editing server device in which the media editing server device is connected to a terminal used by a user via a network, and the media editing server device stores an API for processing media. An API library unit that receives the selection information for acquiring an edit graph from the edit graph storage unit, and a graph selection that acquires the edit graph from the edit graph storage unit based on the received selection information And
A media destination input unit that receives media designation information for designating a media input source and a media output destination from the terminal, and a medium that acquires media from the media input source based on the media designation information received by the media destination input unit A receiving unit, wherein the execution unit acquires an API described in the acquired editing graph from the API library unit, and receives the media based on the acquired editing graph and the acquired API 4. The media editing server device according to claim 1, wherein the media edits the acquired media and outputs the edited media to a media output destination of the media designation information. 5.

In the invention according to claim 5, the terminal inputs the plurality of APIs or the API and a setting value of a parameter of the API to the media editing server device, and the graph selection unit includes the graph selection unit. The edit graph generated by the graph estimation unit is the edit graph to be acquired.
The media editing server device according to claim 4, wherein:

The invention according to claim 6 is a media editing system in which a terminal used by a user and a media editing server device are connected via a network, and the media editing server device includes:
An API for processing media that is video or audio and a basic API that has different API parameter values associated with each other, and a basic API-compatible storage unit that is stored for each API;
A transition probability storage unit in which transition probabilities between the basic APIs are stored, and basic APIs corresponding to the plurality of input APIs from the basic API correspondence storage unit in response to input of a plurality of APIs. A basic API selection unit for reading out and a transition probability between basic APIs input from the basic API selection unit is read from the transition probability storage unit, and the basic API is connected based on the read transition probability A media editing system comprising: a graph estimation unit that generates a graph; and an execution unit that edits media based on the editing graph generated by the graph estimation unit.

  According to the seventh aspect of the present invention, a computer that is a media editing server device stores an API for processing media that is a moving image or audio, and a basic API having different API parameter values for each API. Corresponding to the plurality of input APIs according to the input of a plurality of APIs, a transition probability storage unit storing transition probabilities between the basic APIs, and a plurality of APIs A basic API selection unit that reads and outputs a basic API from the basic API correspondence storage unit, and a transition probability between the basic APIs input from the basic API selection unit is read from the transition probability storage unit, and the read transition probability is A graph estimation unit for generating an edit graph to which the basic API is connected based on the edit graph generated by the graph estimation unit A media editing program to function as an execution unit for editing the media.

According to the present invention, since the template in which the order for editing the media that is a moving image or sound is described is generated by estimation based on the transition probability, the template can be generated.
Thereby, there exists an effect that the user can edit a medium with the produced | generated various templates.

<First Embodiment>
The first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of a video editing system according to an embodiment of the present invention.
As shown in FIG. 1, in the video editing system, a terminal 103A, a terminal 103B, a video source terminal 104, and a video destination terminal 105 are connected to a video editing server 102 via a network 101, respectively.
The terminal 103A, the terminal 103B, the video source terminal 104, the video destination terminal 105, and the video editing server 102 are uniquely identified on the network 101 by identification information such as an IP address.

  The terminal 103A or the terminal 103B is a terminal used by the user, such as a mobile phone, a personal computer, or a portable information device. It is assumed that an input device, a display device, and the like (none of which are shown) are connected to the terminal 103A or 103B as peripheral devices. Here, the input device refers to an input device such as a keyboard and a mouse. The display device refers to a CRT (Cathode Ray Tube), a liquid crystal display device, or the like. Further, the terminal 103A or the terminal 103B has a function of communicating with the video editing server 102 via the network 101.

Here, terminal 103A will be described as a terminal used by a user to edit a video, and terminal 103A will be described as a terminal used by a user to view a video. Further, the terminal 103A and the terminal 103A may be different terminals or the same terminal.
The video source terminal 104 and the video destination terminal 105 have a function of storing video and a function of transmitting and receiving video to and from the video editing server 102 via the network 101.

  Note that the video source terminal 104 and the video destination terminal 105 may be the same terminal or different terminals. The video source terminal 104 and the video destination terminal 105 may be the same terminal as the terminal 103A or the terminal 103B, respectively. The video source terminal 104 and the video destination terminal 105 may each be a moving image storage unit 205 in which videos are stored in the video editing server 102 described later.

Next, the configuration of the video editing server 102 will be described with reference to FIG.
The video editing server 102 stores an API library unit 201 that stores APIs in advance, an editing graph storage unit 203 that stores editing graphs, and a moving image storage unit 205 that stores edited moving images. A ranking DB (database) unit 213 in which a combination of a moving image, an editing graph, and an API, and the number of times of browsing or use are stored, a public browsing unit 206 that ranks and publishes the API, video, and editing graph An edit graph is selected or created, and an edit graph creation unit 202 to which the destination of the movie is input, and an execution unit 204 that edits the movie based on the movie destination and the edit graph received from the edit graph creation unit 202. And having.

Here, the API stored in the API library unit 201 and the edit graph stored in the edit graph storage unit 203 will be described.
The edit graph is information describing a target video, a video sequence, and a video effect processing sequence for the video. Examples of video effect processing include filter processing, composition processing, contrast processing, and rendering processing.

Examples of the filter processing include edge enhancement filter processing, noise removal filter processing, and filter processing for changing a color tone such as monochrome tone or sepia tone.
As the composition process, for example, there are two images, one image is displayed on the entire screen, and another image is displayed on one small screen (hereinafter referred to as a two-screen display process). Alternatively, there is a process of simultaneously displaying a plurality of images on one screen (hereinafter referred to as a multi-screen display process).

Each video effect process has more detailed conditions as parameters. For example, in edge enhancement filter processing, there is a parameter that specifies how many edges are enhanced. Further, for example, in the two-screen display process, there is a parameter that specifies at what position and in what size a small screen is displayed.
An API executes each of these video effect processes, and the API is stored in the API library unit 201 in advance.
In addition, although the image here means an image mainly composed of only an image, it may be an image attached with sound.
The video may be a still image. In the case of a still image, the still image may be handled as a moving image by repeatedly copying the same image.
The video may be a video stored in advance, or a video transmitted and received in real time taken by a WEB camera.

(Description of API)
More specifically, the API used for editing is a single video or a single numerical value or character string, etc., by inputting parameters other than video, such as zero or more videos and zero or more numeric values or character strings. This is to output values other than the video.
Supplementing the output of the API, one or less means the same as allowing a VOID type function in C language or the like. The API is a function, and the API that outputs video does not output any other value. Similarly, an API that outputs a value does not output a video.

Next, typical APIs such as a filter-type API, a composition processing API, a contrast API, and a rendering processing API will be described.
The filter-type API takes one video as input and one filter matrix as a parameter, and outputs the filtered video. Depending on the setting of the filter matrix, it becomes a filter that executes edge enhancement, noise removal (a so-called linear filter), and color conversion (= tone conversion). Note that special noise removal and color conversion (= tone conversion) are often nonlinear filter processing. In addition, an actual function such as OpenCV is often handled separately from the filter process (the function itself is different), but the filter process is handled in a broad sense.

The composition processing API is an API that synthesizes two or more images and outputs one image. There are three types of basic composition processing: layout composition, superposition, and time connection.
The layout composition is a composition in which different videos are arranged and flowed on the same screen, such as multi-dimensional relay. The parameter includes position information on the screen.

Superposition is a combination of processes that change while gradually fading out to another image, such as fade-out. A ratio that determines the degree of pixel value composition (α blending) is passed as a parameter. Note that it is common to pass a blend function as a function pointer as a parameter, as in C language.
The time connection is, for example, a process of playing the video B after the video A ends (for example, the video signal A stops). There are basically no parameters.

  The contrast API is a function that can change the lightness ratio of white (RGB = 255, 255, 255) to black (RGB = 0, 0, 0). Each of the input and output is one image, and the parameter is Ratio (eg 700: 1). Color conversion processing that prevents the color skipping phenomenon in which all the high gradation colors look the same due to the gamma characteristics of a display device such as a display may be included in this type.

  The rendering process API is a process for obtaining a two-dimensional image by arranging a 3D model three-dimensionally, setting a light source and a camera position, and performing perspective projection (a projection method in which a distant object becomes smaller). . There are various types of rendering methods, such as Z-buffer method, ray tracing method, radiosity method, and photon mapping. Basically, the input is not a video but a non-video such as one modeling space, and one video is output. That is, the rendering process is a general term for processes such as converting data other than video into video.

  In addition, there is an API that outputs other than video. For example, there is an API that returns an average value of all pixel values of a video, and an API (for rendering) that returns a shape model by receiving a video.

(Explanation about edit graph)
An edit graph defines a chain structure of APIs as an edit graph in which parameters necessary for the API are given and an output video or value is used as another API input.
Note that the edit graph manages the edit graph as one connection, and therefore an API with completely independent input / output cannot exist. In this case, two different editing graphs are obtained.

Here, the edit graph and the scenario template described in the present embodiment are the same. However, the scenario template is modeled by the user to some extent, and an edit graph that does not generate any meaningful video is not desirable.
Therefore, in the present embodiment, among a number of edit graphs, it is assumed that an edit graph ranked higher in the public browsing described later corresponds to the scenario template. In other words, a scenario template that is frequently used in the editing graph and that is suitable for the user as an example is a scenario template.
In addition, edit graphs obtained by selecting, searching, and estimating edit graphs described later are presented in descending order, and thus the presented edit graphs correspond to scenario templates.

Next, the API stored in the API library unit 201 and the edit graph stored in the edit graph storage unit 203 will be specifically described with reference to FIGS. 3 and 4.
Here, the data structure of an editing graph P510, which is an editing graph as an example, illustrated in FIG. 3A will be described as an example. The edit graph P510 is composed of five objects O1 to O5. One object corresponds to one video process, that is, one API.

The type of API is described by API index information. API index information is also stored in the API library unit 201.
For example, in API index information P520 as shown in FIG. 3B, an identification number for identifying an API (referred to as an API identification number) is described as shown in (a).
Each API of the API index information P520 has an API name, input / output video, and parameter type described in order in association with the API identification number, as shown in FIG. Yes.
The API index information described above serves as a reference of the data structure type and parameter arrangement in the edit graph stored in the edit graph storage unit 203.

FIG. 4 is a diagram showing an API as shown in FIG. 3B stored in the API library unit 201.
In general, each API receives zero or more video streams and zero or more numerical parameters as shown in FIG. 4A and outputs one video stream or one value. For example, FIG. It is a shape like filter P501 of (a).
However, as shown in FIG. 4B, the duplicator P502, which is an API that copies an image and flows it to a plurality, exceptionally has a plurality of outputs. The video stream here represents a continuous flow of image information.

Returning to the description of FIG. 3, based on the API stored in the API library unit 201 described above, the edit graph has a data structure as shown in FIG. The edit graph is stored in the edit graph storage unit 203 as one file.
In the edit graph, first, an object identification number and an API identification number of an API corresponding to the object are described as a set, and an API parameter for each object is described. In FIG. 3C, each row represents one object.
Here, the first line in FIG. 3C corresponds to the object O1 in FIG. 3A, the second line is the object O2, the third line is the object O3, the fourth line is the object O4, 5 The line corresponds to the object O5.

In the edit graph, as shown in (c) and (d), an API object number and an API identification number corresponding to the API object number are described as a set. This is because the same API may be used a plurality of times to distinguish each API.
From this set, for example, the first line of the editing graph in FIG. 3C indicates that the object 1 is an API with API identification number 1 shown in FIG. 3B.

Also, as shown in (e), parameters are described in the order described in the API index information.
Numerical values and character strings are described in the input / output parameters. As a special example of the parameter, there is a case where a matrix is included in the parameter as shown in (f). In addition, as shown in (ki), the parameter may be a character string. When the parameter is a character string, the character string aaa is enclosed in double quotations, such as ““ aaa ””.

  Further, as shown in (c), the input or output from the second parameter of API object number 4 is described by the description of “4 ¥ 2”. In the edit graph, such linkage information indicating the connection of input or output between APIs is described.

  The character strings of (I) input and (K) OutPut are predetermined character strings and indicate that data is input / output from the outside. With this description, for example, the video destination terminal 105, the video source terminal 104, or the moving image storage unit 205 is set as a moving image input source or output destination.

As shown in FIG. 3C, an edit graph includes an API object number, an API identification number corresponding to the API object number, an API parameter identified by the API identification number, and a combination of API objects. Contains information.
Such an edit graph is stored in the edit graph storage unit 203 as a single file with a file name including an edit graph identification number, for example, as shown in FIG.
Note that the edit graph can be described in an XML (Extensible Markup Language) format in addition to the text format illustrated in FIG.

As shown in FIG. 3A, the edit graph is displayed as a diagram based on the link information of the edit graph shown in FIG. 3C and the API parameter type shown in FIG. It is possible.
For example, the public browsing unit 206 or the selection unit 214 described later acquires parameters from the API index information based on the API identification number, and as illustrated in FIG. Can be displayed.

As described above, an identification number (API identification number) for identifying an API is attached to the API. The API is a program or function for executing video processing on a moving image. The API index information describes the type of each API.
The edit graph describes API identification numbers and their order, parameters for each API, and the like. This edit graph is stored in the edit graph storage unit 203 as a file having a file name including an edit graph identification number.
Hereinafter, for the sake of explanation, the API will be described as including an API and an API index.

Next, each configuration of the video editing server 102 in FIG. 2 will be described in detail.
In the API library unit 201, an API for processing a video and an API identification number are stored in advance in association with each other.

  The edit graph storage unit 203 associates an edit graph in which the order of API identification numbers for editing a moving image is described with an edit graph identification number (edit graph identification number) for identifying the edit graph. It is remembered. For example, the edit graph storage unit 203 stores the edit graph as a file including the edit graph identification number in the file name, and searches for the file name including the edit graph identification number using the edit graph identification number as a search key. By doing so, it is possible to search for the file of the corresponding edit graph.

  The moving image storage unit 205 stores the edited moving image and an identification number (moving image identification number) for identifying the edited moving image in association with each other. For example, as with the edit graph storage unit 203, a moving image identification number is included in the file name of the moving image.

The ranking DB unit 213 includes an editing graph identification number of the editing graph stored in the editing graph storage unit 203 and a moving image identification number of the moving image edited using the editing graph stored in the moving image storage unit 205. The relation is stored as a graph moving image correspondence table T1.
In the ranking DB unit 213, the moving image identification number, the number of times of using and browsing the moving image identified by the moving image identification number are associated with each other, and stored as an image number table T2.
In the ranking DB unit 213, the edit graph identification number is associated with the number of times of use and the number of browsing of the edit graph identified by the edit graph identification number, and is stored as the edit graph count table T3. Yes.
In the ranking DB unit 213, the API identification number, the number of times the API identified by the API identification number is used, and the number of browsing times are associated with each other and stored as an API number table T4.

  The number of uses is the number of times each of the edit graph, API, or moving image is used for editing the moving image executed by the execution unit 204. The number of browsing times is the number of times each of the editing graph, API, or moving image is browsed by the user of the terminal 103B via the public browsing unit 206.

  The public browsing unit 206 includes a transmission / reception unit 2061, a ranking acquisition unit 2062, an information acquisition display unit 2063, and a browsing number registration unit 2064 that receive information from the terminal 103B and transmit information for browsing to the terminal 103B.

The ranking acquisition unit 2062 receives an image identification number and an editing graph identification from the image count table T2, the edit graph count table T3, or the API count table T4 of the ranking DB unit 213 in response to a request from the terminal 103B received via the transmission / reception unit 2061. It has a function of acquiring the number or API identification number and the number of times of use or browsing, and summing the obtained number of times of use or browsing and ranking each image, edit graph, or API.
For example, the ranking acquisition unit 2062 acquires the usage count and the browsing count for each video identification number from the image count table T2, sums the usage count and the browsing count for each video identification number, and the video identification number having a large total number Ranking videos in the order of.

  The ranking acquisition unit 2062 associates the edit graph identification number acquired from the graph video correspondence table T1 with the video identification number, or the acquired image identification number based on the edit graph acquired from the edit graph storage unit 203, The editing graph identification number or API identification number has a function of executing ranking by combining the number of times of use or the number of times of browsing.

  For example, the ranking acquisition unit 2062 acquires the usage count and browsing count for each video identification number from the image count table T2, sums the usage count and browsing count for each video identification number, and edit graph count table T3. The number of times of use and the number of times of browsing for each editing graph identification number are obtained from the above, and the number of times of use and the number of times of browsing for each editing graph identification number are summed. The edit graphs identified by the edit graph identification number are ranked.

  Further, the information acquisition / display unit 2063 acquires, from the moving image storage unit 205, the moving image associated with the moving image identification number corresponding to the higher moving image identification number based on the ranking, which is ranked by the ranking acquiring unit 2062. Also, the edit graph associated with the higher-order edit graph identification number is acquired from the moving image storage unit 205, and the edit graph is acquired from the API library unit 201 associated with the higher-order API identification number and transmitted to and received from the terminal 103B. A function of transmitting and displaying via the unit 2061;

  The information acquisition / display unit 2063 has a function of associating and displaying the moving image, the edit graph used for editing the moving image, and the API on the terminal 103B based on the graph moving image correspondence table T1.

The browsing number registration unit 2064 detects that the user has viewed the editing graph, API, or video displayed on the terminal 103B, and the image count table T2 of the ranking DB unit 213 of the editing graph, API, or video viewed by the user, It has a function of updating the browsing count stored in the edit graph count table T3 or the API count table T4.
For example, when the browsing number registration unit 2064 detects that the user of the terminal 103 </ b> B has browsed the moving image, the browsing number associated with the moving image identification number of the detected moving image stored in the image number table T <b> 2. Update.

  For example, when the browsing number registration unit 2064 detects that the user of the terminal 103B has browsed the moving image, the browsing number registration unit 2064 obtains the editing graph identification number corresponding to the moving image identification number of the detected moving image from the graph moving image correspondence table T1. Acquire the edit graph corresponding to the acquired edit graph identification number from the edit graph storage unit 203, and then associate the detected graph with the video identification number of the moving image and store it in the image count table T2. The number of browsing times acquired and related to the editing graph identification number stored in the editing graph number table T3 and the number of browsing times related to the API identification number described in the acquired editing graph are stored in the API number table T4. The number of browsing times being updated is updated.

In the present embodiment, updating means that the number of times of use or the number of times of browsing stored in the image count table T2, the edit graph count table T3, or the API count table T4 is increased and registered. is there.
In the present embodiment, browsing means that the user selects an edit graph, API, or video displayed on the display unit of the terminal 103B using the input unit of the terminal 103B.

The editing graph creation unit 202 includes a transmission / reception unit 2021 for transmitting / receiving to / from the user terminal 103A, a moving image destination input unit 2022, an editing graph selection unit 2025, a graph editing unit 2024, and a graph moving image transmission unit 2023.
The video destination input unit 2022 sends video transmission / reception unit 2021 from the user of the terminal 103A to the video designation information for setting the video to be edited (video source) and the output destination (video destination) of the edited video. It has the function to receive via.

This video designation information identifies, for example, that the input destination of the video to be edited is a video of the video source terminal 104, and that the output destination of the edited video is the video destination terminal 105. Information. For example, it is information such as the IP address, directory name, and file name of the video source terminal 104 or the video destination terminal 105. Note that a moving image registered in the moving image storage unit 205 may be specified as the video designation information.
The information of the video source and video destination received and set here is related to each of the input and output of the editing graph described with reference to FIG. When there are a plurality of video sources and video destinations, a plurality of video sources and video destinations can be designated as input1, input2, and input3.

The graph selection unit 2025 has a function of receiving selection information for selecting an editing graph from the terminal 103A via the transmission / reception unit 2021, and acquiring an editing graph from the editing graph storage unit 203 according to the received selection information. .
The graph selection unit 2025 receives an API identification number or a graph identification number as selection information from the terminal 103A. In this case, the graph selection unit 2025 displays the API or edit graph on the terminal 103A, and selects the API or edit graph on which the user of the terminal 103A is displayed. Here, the displayed API or edit graph is associated with the API identification number or edit graph identification number which is its identification number.
Next, the terminal 103 </ b> A transmits the API identification number or the edit graph identification number selected by the user to the graph selection unit 2025. The graph selection unit 2025 receives the selected API identification number or edited graph identification number as selection information from the terminal 103A.
Alternatively, the user directly inputs the API identification number or the edit graph identification number to the terminal 103A, and the terminal 103A transmits the input API identification number or the edit graph identification number to the graph selection unit 2025. The graph selection unit 2025 receives the API identification number or the edited graph identification number as selection information from the terminal 103A.
The function by which the graph selection unit 2025 selects the edit graph will be described in detail later.

The graph editing unit 2024 acquires an API from the API library unit 201 in response to a user input of the terminal 103A input via the transmission / reception unit 2021, and the user newly creates an editing graph based on the acquired API. It has the function to do.
The graph editing unit 2024 has a function for the user to edit the edited graph acquired by the graph selecting unit 2025. In addition, the graph editing unit 2024 has a function of allowing the user to edit the editing graph by combining the editing graph acquired by the graph selecting unit 2025 and the API acquired by the graph editing unit 2024.

  The graph moving image transmission unit 2023 responds to the video designation information input to the moving image destination input unit 2022 and the user input in response to input of confirmation of editing from the user of the terminal 103A via the transmission / reception unit 2021. The graph selection unit 2025 has a function of transmitting the editing graph acquired by the graph selection unit 2025 or the editing graph created by the graph editing unit 2024 in response to a user input to the execution unit 204.

The execution unit 204 includes a video generation unit 2041, a registration unit 2042, a video transmission unit 2043, and a video reception unit 2044.
The video receiving unit 2044 has a function of acquiring video from the video source based on the video designation information received from the editing graph creating unit 202. For example, the video receiving unit 2044 acquires a video to be edited from the video source terminal 104.

  The video generation unit 2041 has a function of acquiring an API from the API library unit 201 based on the API identification number described in the edit graph received from the edit graph creation unit 202. The video generation unit 2041 has a function of editing video based on the received editing graph, the acquired API, and the video acquired by the video receiving unit 2044.

  The video transmission unit 2043 has a function of transmitting the video edited by the video generation unit 2041 to the video destination based on the video designation information received from the editing graph creation unit 202. For example, the video transmission unit 2043 transmits the edited image to the video destination terminal 105.

The registration unit 2042 has a function of registering the video edited by the video generation unit 2041 in the video storage unit 205 with a video identification number.
The registration unit 2042 has a function of registering the edit graph received from the edit graph creation unit 202 in the edit graph storage unit 203 with an edit graph identification number.
Note that the registration unit 2042 does not register the edit graph when the same edit graph is already registered in the edit graph storage unit 203, and does not register the edit graph already registered in the edit graph storage unit 203. The edit graph identification number is taken as the edit graph identification number.

  Further, the registration unit 2042 has a function of associating the registered moving image identification number with the edit graph identification number used for editing and registering it in the graph moving image correspondence table T1 of the ranking DB unit 213. Also, the registration unit 2042 updates the number of uses of the edit graph identification number used for editing in the edit graph number table T3 of the ranking DB unit 213, and is described in the edit graph used for editing in the API number table T4. It has a function of updating the number of times each API identification number is used.

In addition, when the video source based on the video designation information received from the editing graph creation unit 202 is the video storage unit 205, the registration unit 2042 acquires the video identification number of the video source based on the video designation information from the video storage unit 205. Then, the number of uses of the acquired video identification number in the image count table T2 is updated.
As described above, the registration unit 2042 has a function of updating the number of uses of each of the edit graph, API, and moving image used for editing. In this way, by updating the number of times of use, it is possible to evaluate the moving images registered in the editing graph, API, and moving image storage unit 205 by ranking.

  Next, the operation of the video editing system when the user of the terminal 103B browses will be described with reference to FIG. Here, a case where an edit graph having a higher rank is displayed will be described. First, the user of the terminal 103B transmits a request for browsing to the video editing server 102 (step S501).

The public browsing unit 206 of the video editing server 102 that has received the request from the terminal 103B receives the usage count and the browsing count from the image count table T2, the edit graph count table T3, or the API count table T4 of the ranking DB section 213. . Also, the public browsing unit 206 acquires correspondence information from the graph moving image correspondence table T1 (steps S502 and S503).
Next, the public browsing unit 206 ranks the graph identification numbers based on the acquired usage counts or browsing counts and the correspondence relationship (step S504).

Next, the public browsing unit 206 selects an upper editing graph identification number by ranking, and acquires an editing graph from the editing graph storage unit 203 based on the selected editing graph identification number (steps S505 and S506).
Next, the public browsing unit 206 acquires an API from the API library unit 201 based on the API identification number described in the acquired editing graph (steps S507 and SA508).

  Next, the public browsing unit 206 acquires the moving image identification number associated with the edited graph identification number from the graph moving image correspondence table T1, and uses the acquired moving image identification number and the number of uses and the number of browsing acquired from the image count table T2. Based on the video identification number of the highest ranking, the video identification number of the highest ranking is obtained from the ranking of the obtained video identification number A moving image is acquired from the moving image storage unit 205 (steps S509 and S510).

Next, the public browsing unit 206 transmits to the terminal 103B based on the acquired edit graph, API, and video, and displays the edit graph and the video corresponding to the edit graph as a set on the terminal 103B (step S511). ).
Next, when the public browsing unit 206 detects that the user of the terminal 103B browses, the public browsing unit 206 updates the browsing count stored in the browsed editing graph, API, or video ranking DB unit 213 ( Step S512).

As described above, it is possible to display a combination of an edit graph having a high rank in the ranking of the edit graph and a moving image having a high rank among the videos edited using the edit graph.
FIG. 6 shows a display example as an example that is publicly browsed by the public browsing unit 206 and displayed on the display unit of the terminal 103B. As shown in 6B of FIG. 6, the edit graph and the moving image corresponding to the edit graph are displayed as a set.
Also, as shown in 6A of FIG. 6, APIs are similarly ranked, and it is also possible to display a set of a high API ranking and a video with the highest ranking using the API. .

As shown in FIG. 6, the public browsing unit 206 displays the basic API display part P401 and the recommended template display unit P402 side by side in one screen. The basic APIs are listed in descending order of the number of times they have been used, and processing API names, processing result moving image examples, and types of necessary parameters are displayed as a group.
The example of the moving image as the processing result displayed by the public browsing unit 206 displays the moving image having the largest number of browsing times when the moving image storage unit 205 previously outputs a moving image using only the corresponding API. Further, the public browsing unit 206 displays the scenario template name, the moving image (processing result moving image) processed by the video editing process and the editing graph as a set as to the recommended template.
When displaying the recommended template display unit, the public browsing unit 206 presents the processing result moving image browsing count and the corresponding scenario template usage count in descending order. In addition, when a specific API is designated by the user, the public browsing unit 206 ranks and displays the scenario templates including them.

Next, the operation of the video editing system when the user of the terminal 103A edits a moving image will be described with reference to FIG. Here, a case will be described in which the user of the terminal 103A specifies the video source terminal 104 as the video source, specifies the video destination terminal 105 as the video destination, and selects an edit graph by the edit graph identification number.
First, the user of the terminal 103A transmits a request for editing to the video editing server 102 (step S701).

  The edit graph creation unit 202 of the video editing server 102 that has received the request from the terminal 103A receives an edit graph identification number and input of video designation information from the user of the terminal 103A (step S702).

Next, the edit graph creation unit 202 acquires an edit graph from the edit graph storage unit 203 based on the received edit graph identification number (steps S703 and S704).
Next, the edit graph creation unit 202 transmits the acquired edit graph and the received video designation information to the execution unit 204 (step S705).

The execution unit 204 that has received the editing graph and the video designation information from the editing graph creation unit 202 acquires a video from the video source terminal 104 based on the received video designation information (steps S706 and S707).
Next, the execution unit 204 acquires an API from the API library unit 201 based on the API identification number described in the received edit graph (steps S708 and S709).
Next, the execution unit 204 edits the video based on the acquired video, the acquired API, and the received editing graph (step S710).

Next, the execution unit 204 transmits the edited video to the video destination terminal 105 based on the received video designation information (step S711).
Further, the execution unit 204 registers the edited video with a moving image identification number in the moving image storage unit 205 (step S712).
Further, the execution unit 204 registers the edit graph received from the edit graph creation unit 202 and used for editing, with the edit graph identification number assigned to the edit graph storage unit 203 (step S713).

  Further, the execution unit 204 associates the moving image identification number with the edit graph identification number and registers them in the graph moving image correspondence table T1 of the ranking DB unit 213. The execution unit 204 also stores the moving image identification number, editing graph identification number, and API usage count used for editing stored in the image count table T2, the edit graph count table T3, and the API count table T4 of the ranking DB 213, respectively. Update (step S714)

Next, processing performed in the edit graph creating unit 202 in step S702 in FIG. 7 will be described in detail with reference to FIGS.
First, with reference to FIG. 8, a description will be given of a method in which a user performs editing work by connecting the user to the transmission / reception unit 2021 of the editing graph creation unit 202 through the terminal 103A.
First, the transmission / reception unit 2021 of the editing graph creation unit 202 executes a function of editing when the user is connected via the terminal 103A (step B300), and waits for the user to select an editing function (step B301). ).

The editing function includes an API selection function (step B303) for selecting an API from the API library unit 201, a selection function (step B310) for selecting an edit graph based on ranking, and an edit graph including a required API is searched. Edit graph search function (step B304), edit graph generation function for estimating an edit graph including a required API (step B305), and component assembly function for combining the acquired API and edit graph and changing the value of each parameter (step B302), a video output destination function (step B306) for designating a video output destination, and a video input source designation function (step B307) for designating a video input source.
The user repeatedly performs function selection and inputs information indicating confirmation (step B308).

The transmission / reception unit 2021 executes each function according to the user's function selection. Further, it is detected whether or not an edit graph or video designation information has been input by receiving the confirmation information input (step B308). If not, a warning is displayed on the terminal 103A. Then, the selection function (step B301) is repeated.
The transmission / reception unit 2021 executes an editing process when the detection result of step B308 is input, and transmits the editing graph and video designation information to the execution unit 204 via the graph moving image transmission unit 2023. (Step B309).
Thereafter, the processing described after step S705 in FIG. 7 is performed.

Note that the functions of the video output destination function (step B306) and the video input source designation function (step B307) are the functions of the moving picture destination input unit 2022 described above, and the API selection function (step B303) and the partial assembly function ( Step B302) is a function of the graph editing unit 2024.
The functions of the selection function (step B310), the edit graph search function (step B304), and the edit graph generation function (step B305) are functions of the graph selection unit 2025.

Next, the function of the graph selection unit 2025 will be described in detail. The graph selection unit 2025 includes a graph selection unit 214, a graph search unit 208, and a graph estimation unit 209. Edited by any of the graph selection unit 214, the graph search unit 208, and the graph estimation unit 209 It has a function of acquiring an edit graph from the graph storage unit 203.
The functions of the graph selection unit 214, the graph search unit 208, and the graph estimation unit 209 are the selection function (step B310), the edit graph search function (step B304), and the edit graph generation function (step B305). is there.

The graph selection unit 214 includes a ranking acquisition unit 2141 and an information acquisition display unit 2142. Similar to the public browsing unit 206, the graph selection unit 214 ranks and displays moving images, APIs, or editing graphs on the user terminal 103A. Have
The ranking acquisition unit 2141 has the same function as the ranking acquisition unit 2062 of the public browsing unit 206, and the information acquisition display unit 2142 has the same function as the information acquisition display unit 2063 of the public browsing unit 206. The description of the same function is omitted.

  However, unlike the information acquisition display unit 2063, the information acquisition display unit 2142 receives the selection of the edit graph from the user terminal 103A from the displayed edit graph, and acquires the edit graph identification number of the selected edit graph. And a function of acquiring an edit graph from the edit graph storage unit 203 based on the edit graph identification number of the selected edit graph.

  The selection unit 214 displays the ranked edit graph and the ranked API similar to those in FIG. 6 on the screen of the terminal 103A, and the user selects from the displayed edit graph. In this way, the user can easily select an edit graph or API by viewing the displayed edit graph and its moving image, or API and its moving image.

Next, the graph search unit 208 has a function of receiving an API identification number from the terminal 103 </ b> A and searching for and obtaining an edit graph including the received API identification number from the edit graph storage unit 203.
Here, the API identification number transmitted from the terminal 103A and received by the graph search unit 208 may be a plurality of API identification numbers. Retrieved from the edited graph storage unit 203.

In addition, the graph search unit 208 acquires the number of times of browsing or use of the edited editing graph from the editing graph number table T3 of the ranking DB unit 213, and ranks the edited graph searched based on the acquired number of times of browsing or use. And the edited graph retrieved based on the ranking is displayed on the terminal 103A.
In addition, the graph search unit 208 has a function of receiving a graph identification number from the terminal 103A, and searching and acquiring the edit graph from the edit graph storage unit 203 based on the received edit graph identification number.

Next, the graph estimation unit 209 calculates a connection transition probability graph (= connection probability network (NW)) between APIs based on the editing graph stored in the editing graph storage unit 203, and identifies the API from the terminal 103A. It has a function of generating an edit graph including the received API identification number by receiving the number and searching from the transition probability graph in which the maximum probability path including the received API identification number is calculated.
Here, the graph estimation unit 209 has a function of generating an edit graph with a probability of the edit graph as a whole for each generated edit graph. Therefore, the graph estimation unit 209 can rank the edited graphs generated in descending order of probability. The graph estimation unit 209 has a function of displaying, on the terminal 103A, edit graphs generated in descending order based on the ranking.
In addition, the graph estimation unit 209 displays the generated edit graph on the terminal 103A, and selects the edit graph from the generated edit graph by the user of the terminal 103A selecting the edit graph from the displayed edit graph. Has the function of

  The transition probability graph is stored in a storage unit included in the graph estimation unit 209. The graph estimation unit 209 performs the editing every time an edited graph is registered in the editing graph storage unit 203 or at predetermined time intervals. A transition probability graph is calculated and stored.

The graph state update in the edit graph estimation unit 209 will be described with reference to FIG. Here, a case where the number of APIs stored in the API library unit 201 is 6 from A to F will be described. Also, one API stored in the API library unit 201 will be described in association with one node.
First, as indicated by D601 in FIG. 9A, in the initial state of the transition probability graph, the transition probability to other nodes is 0 in all nodes.

Next, the graph estimation unit 209 has two nodes x and y in all the editing graphs stored in the editing graph storage unit 203, and calculates the number N (x, y) of connections from x to y. And simultaneously calculate the number M (x) of all connections from x to others. At this time, the transition probability P from x to y is calculated by the following equation 1.
P = N (x, y) / (M (x) +1) (Formula 1)
This transition probability P is calculated for each input in each API. That is, in the above example, when there are two inputs 1 and 2 in y, the probability is calculated by counting the number of connections for each input.

With this processing, every time an edit graph as shown in D604 in FIG. 9D and D605 in FIG. 9E is sequentially registered in the edit graph storage unit 203, a transition probability graph is sequentially generated. It is updated in order of D602 in FIG. 9B and D603 in FIG.
As described above, every time an edited graph is registered in the edited graph storage unit 203, the latest transition probability graph is stored in the storage unit of the edited graph estimation unit 209.

  Next, with reference to FIG. 10, the graph estimation unit 209 searches the transition probability graph stored as indicated by D701 in FIG. 10A by the maximum probability path that passes through the condition node input by the user. By doing so, a method of acquiring the edit graph will be described. Note that the transition probability graph D701 in FIG. 10A is the same as the transition probability graph D603 in FIG.

Here, as shown in FIG. 10B, the user sets the start node, end node, and node information that is neither the start point node nor the end point node as the conditions for estimating the edit graph, such as D703 and D704. Suppose that Here, the start point node is node A, the end point node is node F, and the node that is neither the start point nor the end point is node B.
The graph estimation unit 209 searches the starting point node A for route candidates including other condition nodes, searches the editing graphs D705 and D706 as shown in FIGS. 10C and 10D, and edits the editing graphs D705 and D706. If there are a plurality of candidates such as
As a result, when the user gives a necessary API, an edit graph D707 as shown in FIG. 10E having the highest transition probability is estimated from the calculated transition probability graph.

Note that, here, a case has been described in which an edit graph including three APIs in which a start point and an end point are specified is estimated. Therefore, the graph estimation unit 209 generates one estimated edit graph, but a plurality of edit graphs are generated. In such a case, the graph estimation unit 209 displays the edited graph on the user terminal 103A in descending order of the probability that the condition is satisfied.
Next, the user of the terminal 103A selects from the displayed edit graph, and the graph estimation unit 209 receives the selected information, whereby the graph estimation unit 209 selects the received edit graph.

  Next, a method for acquiring an edit graph from a transition probability graph executed by the graph estimation unit 209 will be described with reference to FIG. A set of APIs as a condition is Γc, and a set of all APIs is Γ. Here, the API identification number will be described as a node. That is, the number of nodes in the search condition (number of elements in the set Γc) is n, and the total number of nodes (number of elements in the set Γ) is m.

First, n API identification numbers of API identification numbers X1, X2,..., Xn are input from the terminal 103A as search conditions for the edit graph (step S1101). That is, the number of APIs input as search conditions is n.
Next, the search length l is initialized with a predetermined value L (step S1102). Next, the counter value i is initialized to 1 (step S1103). Next, it is detected whether the counter value i is equal to or less than the number n of APIs input as a search condition (step S1104).

When the detection result of step S1104 is that the counter value i is n or less, the API Xi is selected as the start node S of the probability transition graph (step S1105).
Next, the start point S and the search length l are set in the function F (S, l), and the partial edit graph that becomes a part of the final edit graph and the transition probability graph A probability-provided list Qi (where i is a natural number of 1, 2, 3,... N) having a plurality of pairs of partial edit graph probabilities is acquired and held (step S1106).
Hereinafter, the list Qi in which such a partial graph and its probability are paired is referred to as a partial graph list. The partial graph list is, for example, the graph A shown in D705 of FIG. 10C and its probability 0.3 or the graph B shown in D705 of FIG. 10D and its probability 0.125. The partial graph list includes a part of API identification numbers of the API identification numbers input as search conditions. The function F (S, l) will be described later in detail with reference to FIG.
Returning to the description of FIG. 11A, after step 1106, the counter value i is incremented by 1, and the processing from step S1104 is repeated (step S1107).

  If the counter value i is greater than n, the detection result of step S1104 is a combination of all combinations of graphs that include all the elements of the set Γc, using the acquired partial graph lists Q1, Q2,. A probability (for example, joint probability) is calculated (step S1108). Next, the editing graphs having a high combination probability are displayed in descending order (step S1109).

Next, with reference to FIG. 11B, the processing with the function F (S, l) in step S1106 in FIG. 10 will be described in detail.
First, it is detected whether or not the search length l is greater than 0 (step S1201).
When the detection result in step S1201 is greater than 0, the counter value j is initialized to 1 (step S1202). Next, it is detected whether or not the counter value j is less than or equal to the total number of nodes m (step S1203).

  If the detection result of step S1201 is less than 0, or if the detection result of step S1204 is that the counter value j is greater than the total number of nodes m, the partial graph lists G1, G2,. The partial graph list is collected as one partial graph list, and the combined partial graph list is output as a result (step S1204). The output of step S1204 becomes the subgraph list Qi with probability in step S1106 of FIG.

If the detection value in step S1203 is that the counter value j is less than or equal to the total number of nodes m, the node Aj is selected as an API from those included in the set Γ (step S1205). For example, the node Aj is selected at random from the set Γ.
Next, the transition probability p of transition from the start node S to the node Aj is acquired using the transition probability from the node S to the node Aj of the transition probability graph (step S1206). Next, it is detected whether or not the node Aj is included in the set Γc (whether or not AjεΓc is satisfied) (step S1207).

If Aj is included in the set Γc as a result of the detection in step S1207, the start node S is set to the node Aj. (Step S1208). Next, the subgraph list G ′ is recursively acquired as a result of the function F (S, l−1) (step S1209). Here, the function F (S, l-1) is recursively calculated, and the search length l is set shorter by one each time the function F (S, l-1) is calculated. Note that the search length l is not limited to shortening by one, but may be shortened by a predetermined value.
Next, the node S as the start node is added to the first node of all the subgraph lists in the subgraph list G ′, and the subgraph list Gj obtained by multiplying the total probability by the transition probability p (where j is 1, Is a natural number of 2, 3,... M (step S1210).
Next, the counter value j is incremented by 1, and the processing from step 1204 is repeated (step S1211).
If the detection result in step S1207 indicates that the node Aj is not included in the set Γc, the process in step S1211 is executed.

  As described above, the function F (S, l) collects G1, G2,..., Gm as one partial graph list from the result of step S1203, and returns the combined partial graph list.

  As described with reference to FIG. 11, the graph estimation unit 209 generates an edit graph that includes a conditional API and has a high probability from the set of conditional APIs based on the result of step S1109.

Next, a specific work example in which the graph editing unit 2024 edits the edited graph will be described with reference to FIG.
The graph editing unit 2024 edits or creates an editing graph based on the editing graph selected by the graph selection unit 2025 from the editing graph storage unit 203 or the API registered in the API library unit 201.
Here, a case where the user designates a specific API and the edited graph is acquired by the graph estimation unit 209 will be described.
First, the user inputs an API (step P310), and the graph estimation unit 209 that has received the input API generates an edited graph and displays the generated edited graph on the user's terminal (step P315).

The user adds or corrects an edit graph displayed on the terminal by changing the value of each API parameter or adding or deleting an API using the function of the graph editing unit 2024 (step P320). .
The function of the graph editing unit 2024 allows the user to change the value of each API parameter of the editing graph, add or delete an editing graph by adding or deleting an API, and then display information indicating input / output determination. And sent to the graph editing unit 2024 (step P325).

When the graph editing unit 2024 receives information indicating input / output determination from the user's terminal, the graph editing unit 2024 sets the edited editing graph as an editing graph used for editing.
Note that the graph editing unit 2024 can construct an editing graph by using the API acquired from the API library unit 201 without using the editing graph acquired by the graph selection unit 2025.

Next, an example of an editing screen displayed on the terminal 103A by the graph selection unit 2025 will be described with reference to FIG.
The public browsing function is displayed as a public browsing palette E411 with a small window on the editing mode E410. This palette can be moved freely by dragging and dropping the top bar. APIs and graphs are added by selecting APIs, searching and estimating graphs within this palette, and dragging and dropping the results into the edit mode window. After that, each input and output are connected with lines in the edit mode window, and the edit graph E413 is completed by inputting parameters.

Next, an example using the video editing server according to the present embodiment will be described.
FIG. 14 shows a state where a moving image is uploaded to the video editing server 803. The editor 801 creates an edit graph on the video editing server as shown in a filter graph example 810 as shown in FIG. 14C and sets the output destination in the video editing server 803.
Next, when the video source is set to the terminal 802 and the video is transmitted from the terminal 802 as a stream, a completed output moving image example 805 as shown in FIG. 14B is accumulated in the video editing server 803.
At the same time, since the moving image can be viewed by a viewer 804 other than the editor, the same function as that of the video posting site can be realized. Since the video posting site can be used not only for browsing but also for beginners to share the skills of skilled workers, the posting of beginners can be increased and the circle of community can be expanded compared to the existing video-only posting site.

  Note that the video editing server 803, terminal 802, editor 801, and viewer 804 in FIG. 14 correspond to the video editing server 102, video source terminal 104, terminal 103A, and terminal 103B in FIG. 1, respectively. 1 is the video editing server 803 in FIG.

  FIG. 15A shows multi-point video composition and broadcasting. Specifically, a plurality of games such as baseball are relayed four-dimensionally. In addition, since the video stream does not need to be a live-action video, if a moving image stored in the terminal 2 or a CG video rendered in the terminal 4 is transmitted to the video editing server as a stream, the region is divided into one screen. They are synthesized.

  A filter graph 910 shown in FIG. 15B is a graph corresponding to FIG. 15A, and each terminal and the video editing server always have only two video streams by combining on the video editing server side. Therefore, the addition of communication is smaller than the conventional system that shares all the videos on the terminal side.

FIG. 16A shows a case in which a system for conducting a meeting between persons at a plurality of points in a virtual space is constructed based on FIG. As shown in the filter graph example 1010 shown in FIG. 16B, on the CG virtual space by the terminal 4, the real image of the person by the terminal 1 and the terminal 3 and the conference video material from the terminal 2 are synthesized, and the terminal Deliver to 1 and 3.
By doing so, it is possible to provide an environment in which a sense of being in the same room is shared more than a conference system based on videophone. Such a system is also easily provided by the video editing server according to the present embodiment.

  In the description of the embodiment, it has been described that the ranking DB unit 213 stores the graph moving image correspondence table T1, the image count table T2, the edit graph count table T3, and the API count table T4. The graph video correspondence table T1, the image count table T2, the edit graph count table T3, or the API count table T4 are not included in the API library unit 201, the edit graph storage unit 203, the video storage unit 205, or other storage units. It may be stored.

  According to the first embodiment described above, a dedicated video editing device is not necessary because video editing is performed on the server, and video editing is performed using a scenario template stored in the server. There is an effect that the movie can be easily edited.

  Further, according to the first embodiment described above, the editing graph, the video edited using the editing graph, and the browsing count or usage count of the editing process (API) for executing editing are stored. In addition, ranking is performed in descending order of the number of times of browsing or use, and an editing graph having a high ranking and high evaluation becomes a scenario template, so that an expert does not need to create a scenario template. There is an effect that a moving image can be easily edited using a scenario template generated and stored by a general user.

  In addition, according to the first embodiment described above, a scenario template that meets the user's needs is provided by generating an edit graph including processing required by the user by estimating the generated edit graph. By doing this, there is no need for an expert to create a scenario template, and it is possible to easily edit a moving image using a scenario template generated and stored by a general user.

  Further, according to the first embodiment described above, scenario templates are generated and stored as a by-product accompanying the operation of generating a moving image, so that it is not necessary for an expert to create a scenario template. The user can easily edit the moving image using the scenario template generated and stored.

In addition, according to the first embodiment described above, since all usage histories such as browsing and editing lead to a template generation work, there is an effect that a template candidate is dynamically changed or added. .
Further, according to the first embodiment described above, the template is generated by estimation based on the transition probability graph, so that even a person who is not skilled in video editing can easily edit the video with the generated template. There is an effect that becomes possible.
In addition, since the template generated by estimation may be a new template, the user can easily create novel video content.

<Second Embodiment>
Also in the first embodiment described above, the graph estimation unit 209 generates an edit graph. The graph estimation unit 209 calculates a transition probability graph of connection between APIs based on the editing graph stored in the editing graph storage unit 203, receives the API identification number from the terminal 103A, and determines the received API identification number. An edited graph including the received API identification number was generated by searching from the transition probability graph in which the maximum probability path including the path was calculated.
In contrast to the first embodiment, in the second embodiment, when the graph estimation unit 209 creates an edit graph, the edit is more diverse and the accuracy of the degree of coincidence with the user's needs is high. The video editing server device is capable of creating a graph.

<Principle>
First, the principle of generation of an edit graph in the second embodiment will be described.
In the second embodiment, one API is classified into a plurality of basic APIs by statistically classifying one API for each API based on the value of the parameter of the API.
First, this basic API will be described. The API has a plurality of parameters, but the setting values of the parameters are not completely independent and may be used in association with each other. By classifying the APIs based on the setting values of the associated parameters, even one API can be distinguished as different APIs that are used depending on the parameter setting values, that is, different APIs as basic APIs. is there.

For example, as shown in FIG. 20A, it is assumed that there is APIx that executes a size change process on an input video. In this case, as shown in FIG. 20 (a1), when the set value of the input parameter for this APIx is 1.25, this APIx functions as APIx1 for executing the enlargement process.
On the other hand, as shown in FIG. 20 (a2), when the set value of the input parameter for this APIx is 1.25, this APIx functions as APIx2 for executing the reduction process. In this case, the APIx1 parameter setting value is a setting value for the enlargement ratio.
As described above, even one APIx can function as different APIs, that is, APIx1 and APIx2, depending on parameter setting values, and can be classified as APIs having different functions.

Also, as shown in FIG. 20 (b), when there is an APIy that executes filter processing on the input video, as shown in FIG. 20 (b1), the set value of the input parameter for this APIy is Tb1. The APIy functions as an APIy1 that executes blurring processing.
On the other hand, as shown in FIG. 20 (b2), when the set value of the input parameter for this APIy is Tb2, this APIy functions as APIy2 for executing edge processing.
As described above, even one APIy can be classified as APIs having different functions as different APIs, that is, APIy1 and APIy2, depending on the parameter setting values.

As described with reference to FIG. 20, in the case of APIx for size change processing, which is an API as an example, functionally or semantically, there are two types of basic APIx1 that is enlargement processing and basic APIx2 that is reduction processing. Can be classified into three basic APIs.
As described above, when the edit graph is sufficiently accumulated, the API described in the edit graph is classified by the distribution of the setting values of the API parameters, thereby classifying one API into two basic APIs. It becomes possible.

  Also, the APIy for filter processing whose parameters are matrix operations can also be classified functionally or semantically into blur processing APIy1 and edge processing APIy2. Even in this case, it is possible to classify the APIy functions into two by treating each operation matrix as a vector and dividing it by a cluster of APIy parameter setting value distributions.

As in the above example, API functions can be classified from parameter setting values without knowing the meaning of API processing. The APIs classified according to the parameter setting values in this way will be described as basic APIs. In particular, in the present embodiment, APIs are classified from the usage history of parameter setting values.
In the above description, the case where the API is classified into two has been described. However, as described later, it is also possible to classify into three or more.

Next, the principle of generation of an edit graph when the basic API of FIG. 20 is used will be described with reference to FIG.
In the estimation of the edit graph in the first embodiment, as shown in FIGS. 21A and 21B, when an edit graph using an API including APIx and APIy is generated, the generated edit graph is Since the connection is made only based on the transition probability between APIs without depending on the setting value of the API parameter, as shown in FIG. 21B, one edit in which APIx and APIy are connected. Only graphs are generated.
Therefore, the APIx and APIy parameter setting values may also cause an undesirable connection.

On the other hand, the case of dividing into basic APIs and generating an edit graph from the transition probability will be described with reference to FIGS. 21 (c) and 21 (d).
Here, as shown in FIG. 21C, the connection probability between the enlargement processing APIx1 and the blurring processing APIy1 or the connection probability between the reduction processing APIx2 and the edge processing APIy2 is high, and the others are low.

In this case, even when only APIx and APIy are specified as the APIs to be specified, the first editing graph of the enlargement processing APIx1 and the blurring processing APIy1 as an editing graph with high connection establishment, as shown in FIG. In addition, it is possible to generate the second edit graph of the reduction process APIx2 and the edge process APIy2.
As described above, even when a small number of APIs are specified, an edit graph is generated in consideration of connections between a plurality of basic APIs into which the specified APIs are classified, so that a plurality of various edit graphs are generated. Can be generated.
In addition, since the transition probability is calculated in consideration of the set value of the API parameter, there is an effect that there is a high possibility that a desirable connection is generated.

If the user specifies the APIx parameter setting value as 1.25 as shown in FIG. 20, that is, specifies the enlargement processing APIx1 and does not specify the APIy parameter setting value, FIG. Based on the transition probability of 21 (c), the APIy connected to the enlargement processing APIx1 generates an editing graph of the enlargement processing APIx1 and the blurring processing APIy1 because the connection of the blurring processing APIy1 is higher than that of the edge processing APIy2.
As described above, in the second embodiment, it is possible to generate an edit graph that more closely matches the user's wishes by generating an edit graph based on the transition probability of not only the API but also the basic API. It is.

Next, the configuration of the video editing system according to the second embodiment will be described with reference to FIG. In the figure, parts corresponding to those in FIG.
In FIG. 17, a basic API extraction server 106 is connected to the video editing server 102 via the network 101.
The basic API extraction server 106 classifies the API described in the editing graph into the basic API based on the setting value of the API parameter for the editing graph stored in the video editing server 102, and classifies it. The transition probability between the basic APIs is calculated.
Further, the basic API extraction server 106 stores the calculated transition probability between the basic APIs in the video editing server 102.
The video editing server 102 classifies the input API into basic APIs according to a plurality of input APIs, and edits the editing graph including the classified basic APIs based on transition probabilities between the basic APIs stored therein. The video is edited and output based on the generated editing graph.

Next, the configuration of the video editing system in FIG. 17 will be described in detail with reference to FIG. In the figure, the same reference numerals are given to portions corresponding to the respective portions in FIG.
The basic API extraction server 106 includes a basic processing classification unit 231, a basic API selection unit 232, a basic API transition probability calculation unit 233, and a classification information storage unit 234.
The video editing server 102 includes an editing graph storage unit 203, a transition probability storage unit 215, and a graph estimation unit 210.

Similar to the first embodiment, the edit graph storage unit 203 stores an edit graph in which the order of API identification numbers for editing a moving image and the set values of a plurality of parameters are described.
The transition probability storage unit 215 stores transition probabilities between basic APIs.
The classification information storage unit 234 stores classification information for classifying setting values of a plurality of parameters used for classifying an API into a basic API.
In the basic API correspondence storage unit 235, an API for video processing of a moving image and a basic API having different API parameter values are associated with each other and stored for each API.

The basic API classification unit 231 reads the edit graph stored in the edit graph storage unit 203, and for each API used in the read edit graph, an API is determined based on setting values of a plurality of parameters of the API. The basic APIs having statistically different parameter setting values are classified, and the APIs and the classified basic APIs are associated with each other and stored in the basic API correspondence storage unit 235.
In addition, when classifying an API into a basic API, the basic API classification unit 231 calculates classification information for statistically classifying setting values of a plurality of parameters, and converts the API into a basic API based on the calculated classification information. The classified and calculated classification information is stored in the classification information storage unit 234.

  The basic API transition probability calculation unit 233 calculates the transition probability between the basic APIs classified by the basic API classification unit 231 based on the transition frequency between the basic APIs of the edit graph stored in the edit graph storage unit 203. The calculated transition probability is stored in the basic API transition probability storage unit 215.

In response to the input of a plurality of APIs, the basic API selection unit 232 reads out the basic API corresponding to the input plurality of APIs from the basic API correspondence storage unit 235, and the graph estimation unit 210 Output to.
The basic API selection unit 232 receives an API identification number and setting values of a plurality of parameters of the API identified by the API identification number from the terminal 103 </ b> A used by the user via the video editing server 102. Accordingly, the classification information is read out from the classification information storage unit 234, and the basic API correspondence storage unit 235 corresponding to the plurality of input APIs based on the read out classification information and the set values of the plurality of parameters input. The basic API corresponding to the input set values of the plurality of parameters is selected from the basic API read out from, and the selected basic API is output to the graph estimation unit 210.

The graph estimation unit 203 reads the transition probability between the basic APIs input from the basic API selection unit 232 from the transition probability storage unit 215, and generates an edit graph to which the basic API is connected based on the read transition probability.
Further, the graph estimation unit 203 reads the transition probability between the basic APIs input from the basic API selection unit 232 from the transition probability storage unit 215, and the basic APIs are connected in the transition order in which the read transition probability is the maximum. Generate an edit graph.

The graph estimation unit 203 generates the edit graph to which the basic API is connected in the order of transition in which the transition probability that maximizes the transition probability is maximized, so that the edit graph stored in the edit graph storage unit 203 is assembled. It is possible to generate a basic API having a transition order with the highest use frequency and an edit graph having the order in the basic API.
Conversely, when the graph estimation unit 203 generates an edit graph with a low transition probability, there is a possibility of generating an edit graph with high originality that is not frequently used.
The edit graph creation unit 202 edits the video based on the edit graph generated by the graph estimation unit 210 and outputs the edited video.

  Compared with the graph estimation unit 209 in the first embodiment, the graph estimation unit 209 in the first embodiment generates an edit graph based on the transition probability graph for the API in the second embodiment. On the other hand, the graph estimation unit 210 in the second embodiment is different in that an edit graph based on a transition probability graph is generated for the basic API.

In the second embodiment, since the estimation unit 210 generates an edit graph based on the transition probability for the basic API, the transition probability between the basic APIs is stored in the transition probability storage unit 215. Yes.
That is, in the first embodiment, the transition probability graph is stored in the storage unit included in the graph estimation unit 209. However, in the second embodiment, the transition probability graph is stored in the transition probability storage unit 215. Is different.

In the graph estimation unit 209, there are two nodes x and y in all the editing graphs stored in the editing graph storage unit 203, and the number N (x, y) of connections from x to y is calculated. At the same time, the number M (x) of all connections from x to the other is calculated, and at this time, the transition probability P from x to y is calculated by the following equation 1. In all the editing graphs accumulated in the accumulation unit 203, there is an output x of a certain API α and an input y of an API β different from a certain API α, and the number N (x, y) of connections from the output x to the input y is expressed as follows. At the same time, the number M (x) of all the connections from x to the other is calculated. At this time, the transition probability P from x to y is calculated by the following equation 1.
In other functions, the graph estimation unit 209 and the graph estimation unit 210 are the same.

Next, operations of the basic API classification unit 231, the basic API selection unit 232, and the graph estimation unit 210 will be described with reference to FIG.
(Step S1901)
First, the basic API classifying unit 231 reads all edit graphs having a certain API from the edit graph storage unit 203 as shown in FIG.
Here, the description will be made assuming that there are m descriptions corresponding to a certain API in the read edit graph. Here, an API is an API having n numerical parameters, and this API is APIf. APIf has parameters from X ₀ to X _n−1 , and APIf = f (X ₀ , X ₁ ,..., X _n−1 ). The function f is a processing function that defines an API.

All editing graph APIf editing graph storage unit 203 is read out is written, the set value of X _n-1 of the parameter from the X ₀ is a parameter APIf each is described. For this APIf, a set of parameter setting values i described in the editing graph is a vector vi. Here, i is an identifier for identifying APIf in the edit graph, and is a natural number.
The edit graph accumulating unit 203 reads the setting values of the APIf parameters, that is, the vectors v1, v2, v3,. The edit graph accumulating unit 203 executes the following principal component analysis with the vectors v1, v2, v3,.

(Step S1902)
Next, as shown in FIG. 19 (2), the basic API classifying unit 231 performs principal component analysis on the input vector set V to obtain λ1 as the first principal component and λ2 as the second principal component. ,calculate.
Here, the principal component analysis is a statistical method of combining (compressing) several correlated factors into several components and obtaining their total power and characteristics. Since the principal component analysis is a general method of statistical processing, the description thereof is omitted here.

Next, the basic API classifying unit 231 calculates the eigenvector of the first principal component λ1 calculated by the principal component analysis as the eigenvector P1, and calculates the average vector meanV for the input vector set V. The eigenvector P1 is a vector whose direction is not changed by linear transformation with respect to the first principal component λ1. The average vector meanV is a vector whose elements have an average value of vector elements, and the average vector meanV when the number of vectors is m is calculated by the following equation 2.
meanV = (v1 + v2 + v3 +... + vm) / m (Expression 2)

The basic API classification unit 231 uses the eigenvector P1 as the first classification information, the calculated average vector meanV as the second classification information, and the first classification information and the second classification information as a set of classification information. Store in the storage unit 234.
As will be described later, since the process is executed recursively, the basic API classification unit 231 stores the first classification information and the second classification information as a pair in the classification information storage unit 234 in order. .
The average vector meanV is obtained by adding the vectors v1, v2, v3,... Vm, which are the input vector set V, and dividing the added vector by the number m of elements of the vector set V.

(Step S1903)
Next, as shown in FIG. 19 (3), the basic API classifying unit 231 has a value (λ1 / λ2) obtained by dividing the first principal component λ1 by the second principal component λ2 than the preset threshold value T. When it is large (λ1 / λ2> T), the input vector V is classified into two set vectors V + and V− determined by the eigenvector P1.
Conversely, the basic API classification unit 231 performs classification when the value obtained by dividing the first principal component λ1 by the second principal component λ2 (λ1 / λ2) is equal to or less than a preset threshold T (λ1 / λ2 ≦ T). To finish.

The set vectors V + and V- are set vectors of one and the other with respect to a plane defined by the eigenvector P1 with the average vector meanV as the origin. Note that the set vector V + is a vector set of a region having a direction that coincides with the direction of the eigenvector P1 with the plane as a reference, and conversely, the set vector V− has the direction of the eigenvector P1 with respect to the plane. It is a vector set of regions in the opposite direction.
When the value obtained by dividing the first principal component λ1 by the second principal component λ2 (λ1 / λ2) is larger than a preset threshold T, the first principal component λ1 is greater than the second principal component λ2. Therefore, it has a significant meaning as a statistical explanatory variable. Alternatively, it means that the distribution of the vector set along the direction of the first principal component λ1 is flatter than the distribution of the vector set along the second principal component λ2. The preset threshold T is a value greater than 1.

  Specifically, the basic API classifying unit 231 calculates an inner product of a vector vi of the input vector V (i is a natural number from 1 to m) and a vector obtained by subtracting the average vector meanV1 from the vector vi and an eigenvector P1. If the calculated inner product value is positive, the vector vi is classified into the vector set V +. Conversely, if the calculated inner product value is not positive, the vector vi is changed to the vector set V−. Classify.

  Next, the basic API classification unit 231 similarly applies the eigenvector P1 that is the first classification information and the second classification information to all the vectors vi (i is a natural number from 1 to m) in the input vector V. Is classified into two vector sets, V + and V−.

As described above, the basic API classifying unit 231 classifies m vectors vi into two vectors V + and V− as shown in FIG. 19 (3).
In FIG. 19 (3), for the sake of explanation, the vector distribution in the three-dimensional space is shown divided into two-dimensional planes, but in general, the vector in the n-dimensional space is illustrated. Are divided and classified by an (n−1) -dimensional hyperplane.

(Step S1904)
Next, the basic API classifying unit 231 recursively repeats the processing from step S1901 to step S1903 for each of the divided vector sets V + and V−, and each vector set. On the other hand, each vector set is classified into two until the value (λ1 / λ2) obtained by dividing the first principal component λ1 by the second principal component λ2 becomes smaller than a preset threshold value T.

As described above, the basic API classifying unit 231 generates a binary tree as shown in FIG.
In FIG. 19 (4), for example, the vector set a0 that is the set value of the APIA parameter is classified into the vector set a1 and the vector set a2, the vector set a1 is classified into the vector set a11 and the vector set a12, The vector set a2 is classified into a vector set a21 and a vector set a22, and the vector set a21 is classified into a vector set a211 and a vector set a212.

That is, one vector set a0 is classified into five vector sets: a vector set a11, a vector set a12, a vector set a211, a vector set a212, and a vector set a22.
This one vector set a0 is an API, and five vector sets of a vector set a11, a vector set a12, a vector set a211, a vector set a212, and a vector set a22 are basic APIs. Hereinafter, for example, a basic API whose parameters are classified as a1 in APIA will be referred to as basic APIa1.
Note that the first classification information and the second classification information are stored in the classification information storage unit 234 as a set corresponding to each classification of FIG. 19 (4).

The basic API classification unit 231 associates the API with a plurality of basic APIs obtained by classifying the API into a binary tree, and stores them in the basic API correspondence storage unit 235.
Further, the basic API classifying unit 231 calculates an average value of the parameters for the plurality of basic APIs classified into the two trees as an average parameter value, and the plurality of basic APIs classified into the two trees. The calculated average parameter value is stored in the basic API correspondence storage unit 235 in association with the calculated average parameter value.

(Step S1905)
Using the basic APIs classified as described above, as shown in FIG. 19 (5), the basic API transition probability calculation unit 233 is based on the frequency of the edit graph stored in the edit graph storage unit 203. The transition probability between the basic APIs is calculated, and the calculated transition probability between the main APIs is stored in the transition probability storage unit 215.
For example, as shown in FIG. 19 (5), APIAs are classified into three types: basic APIa1, basic APIa2, and basic APIa2. Similarly, APIB is classified into two, basic APIb1 and basic APIb2, and APIC is classified into two, basic APIc1 and basic APIc2.

Next, the basic API selection unit 232 sets the transition probabilities between the basic APIs classified by the basic API classifying unit 231 as illustrated in FIG. And is stored in the transition probability storage unit 215.
Next, the graph estimation unit 210 reads the transition probability between basic APIs from the transition probability storage unit 215 and generates an edit graph based on the transition probability between APIs described with reference to FIGS. 9 to 12. Similarly to the above, an edit graph is generated based on the transition probability between the read basic APIs.
The generation of the edit graph based on the transition probability in the first embodiment and the generation of the edit graph based on the transition probability in the second embodiment are based on the transition probability between APIs and the API in the first embodiment. Whereas the edit graph is generated, the second embodiment is different in that the transition probability between basic APIs and the edit graph generated by the API are generated. Other processing is the same.

Next, specific processing of the basic API selection unit 232 will be described.
The basic API selection unit 232 receives, via the video editing server 102, the API identification number and the setting values of a plurality of parameters of the API identified by the API identification number from the terminal 103A used by the user. Accordingly, the classification information is read from the classification information storage unit 234, and the API corresponding to the API identification number classified by the basic API classification unit 231 is classified based on the read classification information and the set values of the plurality of parameters input. The basic API corresponding to the input set values of the plurality of parameters is selected from the input basic APIs.

For example, a case will be described in which the user inputs a parameter setting value and generates an edit graph for the APIA shown in FIG.
In this case, the basic API selection unit 232 classifies the APIA in which the parameter setting value is set to which of the basic APIs based on the first classification information and the second classification information that are classification information. To do. Here, the APIA before classification is APIa0.

(Step S2001)
First, the basic API selection unit 232 sequentially selects the eigenvector P1 that is the first classification information stored as a set and the average vector meanV that is the second classification information from the classification information storage unit 234 in order from the APIA. read out.

(Step S2002)
Next, the basic API selection unit 232 calculates an inner product R of a vector obtained by subtracting the average vector meanV1 from the vector Q and an eigenvector P1 with respect to the input vector Q, which is a parameter setting value, by the following equation 3. When the calculated inner product R is positive, that is, when the calculated inner product R is larger than 0 (R> 0), the vector vi is classified into the vector set V +, and conversely, the calculated inner product R When the value is not positive, that is, when the calculated inner product R is 0 or less (R ≦ 0), the vector vi is classified into the vector set V−.
R = P1 · (Q−meanV1) (Formula 3)
Based on this classification, the basic API selection unit 232 detects whether the APIA (= a0) in which the parameter setting value is set is the basic APIa1 or the basic APIa2.

(Step S2003)
The basic API selection unit 232 repeats steps S2001 and S2002, and sequentially stores the eigenvector P1 that is the first classification information stored as a set and the average vector meanV that is the second classification information, in the classification information storage. The user sequentially classifies the input vector Q, which is a parameter setting value for the APIA, until the first classification information and the second classification information are sequentially read from the unit 234 and stored as a set.

  As described above, the basic API selection unit 232 classifies the input vector Q, which is a parameter setting value for the APIA, into any one of APIa11, APIa12, APIa211, APIa212, and APIa22.

Further, for example, when APIs are classified as shown in FIG. 19 (5), the user sets parameter setting values for APIA, sets parameter setting values for APIB, and sets APIC. On the other hand, when generating the edit graph by setting the parameter setting values, the basic API selection unit 232 classifies the basics that are classified into the APIs APIA, APIB, and APIC that are input. A corresponding basic API is selected from the APIs. For example, APIa1, APIb2, and APIc2 are selected.
Next, the graph estimation unit 210 generates an edit graph including each selected basic API so that the transition probability of each selected basic API is maximized. For example, an edit graph including APIa1, APIb2, and APIc2 is generated.

  Also, for example, when APIs are classified as shown in FIG. 19 (5), the user sets parameter setting values for APIA and does not set parameter setting values for APIB and APIC. In the case of generating an edit graph, the basic API selection unit 232 selects only the APIA for which the parameter setting value is set, based on the set parameter setting value. The corresponding basic API is selected from. For example, APIa1 is selected. Since the basic API selection unit 232 does not set the parameter setting values for APIB and APIC, APIb1 and APIb2 that are APIB basic APIs, and APIc1 and APIc2 that are APIIC basic APIs are set. select.

Next, the graph estimation unit 210 calculates the selected basic API and the basic API of the input API so that the transition probability between the selected basic API and the basic API of the input API is maximized. Generate an edit graph containing.
For example, an edit graph including the selected APIa1, APIb1 or APIb2 that is a basic API of APIB, and APIc1 or APIc2 that is a basic API of APIC is generated.

In generating the edit graph according to the second embodiment described above, since the transition probability with each API as a node represents the probability of transition in relation to the parameter value, the diversity is higher. In addition, there is an effect that it is possible to create an edit graph with high accuracy of the degree of coincidence with the user's needs.
Further, when generating an edit graph according to the second embodiment, even when the number of edit graphs stored in the edit graph storage unit 203 is small, APIs are classified into basic APIs to generate an edit graph. Therefore, it is possible to create an edit graph that is more diverse and that has a high degree of matching with the user's needs.

The advantages of generating the edit graph in the second embodiment when compared with the generation of the edit graph in the first embodiment will be described.
In the generation of the edit graph in the first embodiment, if the API is the same, the transition probability graph is constructed as the same node regardless of whether the parameter setting value is input. There was a case where an edit graph was presented.

On the other hand, in generating an edit graph in the second embodiment, by distinguishing APIs by setting values of API parameters, the media connection template is suitable and the media expression template meeting the user's needs is provided. Becomes a competence.
Furthermore, in the generation of the edit graph in the second embodiment, APIs are classified according to parameter setting values even with a single API, so that variations in APIs increase, and variations in edit graphs using APIs. Has the effect of increasing. This increase in the number of types of APIs and edit graphs has the advantage that the selection range of edit graphs in video editing is widened for the user, and the user can edit various videos.

Next, a mode for realizing the video editing server 102 and the basic API extraction server 106 will be described with reference to FIG.
Form 1 shown in FIG. 22A is a form when the function of the basic API extraction server 106 is integrated into the video editing server 102. As a configuration in this case, (1) a configuration in which the video editing server 102 and the basic API extraction server 106 are different servers, and (2) a configuration in which the video editing server 102 and the basic API extraction server 106 are integrated. It may be possible.
In form 1 shown in FIG. 22A, the video editing server 102 can integrate editing graphs generated by various users and present templates to the users.

On the other hand, form 2 in FIG. 22B is a configuration in which the user terminal 103A or 103B has the function of the basic API extraction server 106.
In form 2 shown in FIG. 22B, the generation of the edit graph is limited to the user who uses the terminal, but it is effective when the user's personal video editing know-how is not desired to be disclosed to the outside.
In either case of form 1 or form 2, since the user who uses the video editing server 102 to edit the video and uses the terminal 103A or 103B does not need a device for editing the dedicated video, It is possible to reduce the cost for an apparatus for editing video.

Although the graph estimation unit 210 in the second embodiment of FIG. 18 has been described as a configuration different from the graph estimation unit 209 in the first embodiment of FIG. 2, the graph estimation unit 210 and the graph estimation unit 209 are The same configuration may be used.
In addition, the video editing server 102 and the basic API extraction server 106 may be configured separately or may be integrated.

  The API library unit 201, the edit graph storage unit 203, the moving image storage unit 205, the storage unit of the ranking DB unit 213, the transition probability storage unit 215, or the classification information storage unit 234 may be a hard disk device, a magneto-optical disk device, a flash A nonvolatile memory such as a memory, a storage medium that can only be read such as a CR-ROM, a volatile memory such as a RAM (Random Access Memory), or a combination thereof is used.

The public browsing unit 206, the edit graph creation unit 202 or the execution unit 204 in FIG. 1, or the graph estimation unit 210, the basic API classification unit 231, the basic API selection unit 232, or the basic API transition probability calculation unit 233 in FIG. It may be realized by dedicated hardware, or may be realized by a memory and a microprocessor.
The public browsing unit 206, the edit graph creation unit 202, the execution unit 204, the graph estimation unit 210, the basic API classification unit 231, the basic API selection unit 232, or the basic API transition probability calculation unit 233 are realized by dedicated hardware. The public browsing unit 206, the edit graph creation unit 202, the execution unit 204, the graph estimation unit 210, the basic API classification unit 231, the basic API selection unit 232, or the basic API transition probability calculation unit. 233 may be configured by a memory and a CPU (central processing unit), and may be implemented by loading a program for realizing the function into the memory and executing the program.

  Further, it is assumed that an input device, a display device, and the like (none of them are shown) are connected to the terminals 103A and 103B as peripheral devices. Here, the input device refers to an input device such as a keyboard and a mouse. The display device refers to a CRT (Cathode Ray Tube), a liquid crystal display device, or the like.

In the description of the above-described embodiment, only the case of a video including a moving image and audio as the processing target of the video editing server and the basic API extraction server according to the present embodiment has been described. The same can be applied to the case where only the moving image, only the sound, or the moving image and the sound are used as the media to be processed by the video editing server and the basic API extraction server.

The media processing server or the media processing server device of the present invention corresponds to the video processing server 102 of the embodiment, and processes the processing on the media.
In the correspondence between the present invention and the embodiment, the media destination input unit is the video destination input unit, the media receiver is the video receiver, the media input source is the video input source, the media output destination is the video output destination, The media designation information corresponds to the video designation information.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design and the like within a scope not departing from the gist of the present invention.

  The present invention is suitable for use in a server device on a network for editing video.

It is a block diagram which shows the structure of the video editing server system by one Embodiment of this invention. It is a block diagram which shows the structure of the video editing server apparatus of FIG. It is explanatory drawing explaining API and an edit graph as an example. It is explanatory drawing explaining the example which shows API in a figure. It is a flowchart which shows the operation | movement at the time of browsing of a video editing server system. It is a screen figure of the browsing display as an example displayed on a user's terminal. It is a flowchart which shows the operation | movement at the time of edit of a video editing server system. It is a flowchart explaining step S702 of FIG. 7 in detail. It is explanatory drawing explaining the production | generation method of a transition probability graph. It is explanatory drawing explaining the method to detect the transition with high probability from the produced | generated transition probability graph. It is a flowchart of the method of detecting the transition with high probability from the produced | generated transition probability graph. It is explanatory drawing for demonstrating edit. It is explanatory drawing for demonstrating graph selection. It is explanatory drawing explaining the 1st application example of this video editing server system. It is explanatory drawing explaining the 2nd application example of this video editing server system. It is explanatory drawing explaining the 3rd application example of this video editing server system. It is a block diagram which shows the structure of the video editing server system by 2nd Embodiment of this invention. It is a block diagram which shows the structure of the video editing server apparatus of FIG. 17, and an extraction server. It is explanatory drawing explaining estimation of the edit graph by 2nd Embodiment. It is explanatory drawing about the classification | category of API. It is explanatory drawing explaining estimation of the edit graph by the classified API. It is a block diagram which shows the example of a structure of a video editing server system.

Explanation of symbols

101 Network 102 Video Editing Server 103A, 103B Terminal 104 Video Source Terminal 105 Video Destination Terminal 106 Basic API Extraction Server 201 API Library Unit 202 Edit Graph Creation Unit 203 Edit Graph Storage Unit 204 Execution Unit 205 Video Storage Unit 208 Graph Search Unit 209, 210 Estimator 213 Ranking DB Unit 214 Selection Unit 215 Transition Probability Storage Unit 231 Basic API Classification Unit 232 Basic API Selection Unit 233 Basic API Transition Probability Calculation Unit 234 Classification Information Storage Unit 2021 Transmission / Reception Unit 2022 Video Destination Input Unit 2023 Graph Video Transmission Unit 2024 Graph editing unit 2025 Graph selection unit 2041 Video generation unit 2042 Registration unit 2043 Video transmission unit 2044 Video reception unit 2061 Transmission / reception unit 2062 Ranking acquisition unit 2063 Information acquisition Display unit 2064 Views registration unit 2141 Index acquiring unit 2142 information acquisition display section T1 graph video correspondence table T2 image number table T3 edit graph count table T4 API count table

Claims (7)

An API for processing media that is video or audio and a basic API that has different API parameter values associated with each other, and a basic API-compatible storage unit that is stored for each API;
A transition probability storage unit in which transition probabilities between the basic APIs are stored;
A basic API selection unit that reads out and outputs a basic API corresponding to the plurality of input APIs from the basic API corresponding storage unit in response to input of a plurality of APIs;
A graph estimation unit that reads a transition probability between basic APIs input from the basic API selection unit from the transition probability storage unit and generates an edit graph to which the basic API is connected based on the read transition probability;
An execution unit for editing the media based on the edit graph generated by the graph estimation unit;
A media editing server device characterized by comprising: The media editing server device
An edit graph storage unit storing an edit graph in which the API order and API parameter setting values are described;
Based on the edit graph stored in the edit graph accumulating unit, for each API, the API is classified into the plurality of basic APIs having statistically different parameter setting values, and the API and the classified basics A basic API classifying unit that associates an API and stores it in the basic API-compatible storage unit;
A basic API for calculating a transition probability between basic APIs classified by the basic API classifying unit based on an editing graph stored in the editing graph storage unit and storing the calculated transition probability in the transition probability storage unit A transition probability calculation unit;
The media editing server device according to claim 1, further comprising: The media editing server device
A classification information storage unit storing classification information for classifying setting values of a plurality of parameters used for classifying an API into a basic API;
The basic API classification unit is
When classifying the API into the basic API, classification information for statistically classifying the setting values of the plurality of parameters is calculated based on the editing graph stored in the editing graph storage unit, and the calculation is performed. Classifying the API into the basic API based on classification information, storing the calculated classification information in the classification information storage unit;
The basic API selection unit
In response to the input of the API and the set value of the API parameter, the classification information corresponding to the input API is read from the classification information storage unit, and the read classification information and the input parameter The basic API corresponding to the input values of the plurality of parameters is selected from the basic APIs read out from the basic API correspondence storage unit corresponding to the input APIs. , Output the selected basic API,
The media editing server device according to claim 2, wherein: The media editing server device
A media editing server device connected to a terminal used by a user via a network,
The media editing server device
An API library section in which an API for processing media is stored;
Receiving a selection information for acquiring an edit graph from the edit graph storage unit from the terminal, and a graph selection unit for acquiring an edit graph from the edit graph storage unit based on the received selection information;
A media destination input unit for receiving media designation information for designating a media input source and a media output destination from the terminal;
A media receiving unit that acquires media from a media input source based on the media designation information received by the media destination input unit;
Have
The execution unit is
The API described in the acquired editing graph is acquired from the API library unit, the media acquired by the media receiving unit is edited based on the acquired editing graph and the acquired API, and the edited media To the media output destination of the media designation information,
The media editing server device according to claim 1, wherein the media editing server device is a device. The terminal inputs the plurality of APIs, or the API and a parameter setting value of the API, to the media editing server device,
The graph selection unit, the edit graph generated by the graph estimation unit is the edit graph to be acquired,
The media editing server device according to claim 4, wherein: A media editing system in which a terminal used by a user and a media editing server device are connected via a network;
The media editing server device
An API for processing media that is video or audio and a basic API that has different API parameter values associated with each other, and a basic API-compatible storage unit that is stored for each API;
A transition probability storage unit in which transition probabilities between the basic APIs are stored;
A basic API selection unit that reads out and outputs a basic API corresponding to the plurality of input APIs from the basic API corresponding storage unit in response to input of a plurality of APIs;
A graph estimation unit that reads a transition probability between basic APIs input from the basic API selection unit from the transition probability storage unit and generates an edit graph to which the basic API is connected based on the read transition probability;
An execution unit for editing the media based on the edit graph generated by the graph estimation unit;
A media editing system comprising: A computer that is a media editing server device
An API for processing media that is video or audio and a basic API that has different API parameter values associated with each other, and a basic API-compatible storage unit that is stored for each API;
A transition probability storage unit in which transition probabilities between the basic APIs are stored;
A basic API selection unit that reads out and outputs a basic API corresponding to the plurality of input APIs from the basic API corresponding storage unit in response to input of a plurality of APIs;
A graph estimation unit that reads a transition probability between basic APIs input from the basic API selection unit from the transition probability storage unit and generates an edit graph to which the basic API is connected based on the read transition probability;
An execution unit for editing the media based on the edit graph generated by the graph estimation unit;
Media editing program to function as

Images