JPH11126205A - Multimedia data presentation system and storage medium stored with program realizing it - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system so as to take out necessary parts of multimedia data stored in a database as many as needed, put them together within a required time and reproduce them as a digest. SOLUTION: A retrieving part 22 retrieves a content object or a view object of a database in a target object storing part 13 by a keyword, etc., and produces an object for execution. A digest data generating part 23 uses data in order of priority that is given to the object for execution and produces digest data. A data displaying part 24 reproduces a part of corresponding multimedia data on a display device 26 based on the produced digest data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a multimedia data presentation system capable of automatically generating a digest for multimedia data, and a storage medium for storing a program for realizing the system.

[0002]

2. Description of the Related Art Multimedia data presentation systems include a system using an object-oriented database. This system enables a flexible presentation such as accessing a part of multimedia data through an object. I have.

On the other hand, regarding the generation of a digest in which only important parts of multimedia data are extracted and summarized, conventionally, an important part of multimedia data is manually cut and pasted using an authoring tool and edited. There was no other way but to do so.

[0004]

[0005] Multimedia data including video data has a parameter of time, and it takes time to browse. Especially when there is a large amount of data, it takes a considerable time to browse, and it becomes impossible to refer to virtually all data.

[0005] As a method of shortening the time for referring to the multimedia data, there is a method of reproducing the data by fast-forwarding. However, there is a limit to the speed that can be recognized by human eyes, and the speed is reduced within a certain time. It is not possible. In order to refer to the multimedia data within a limited time, a digest function of extracting and referencing only a necessary and important part of the multimedia data is required.

However, as described above, in the digest generation method by cutting and pasting using a conventional authoring tool, the necessary data portions are cut out and selected manually, which takes considerable time and effort in the generation processing. Will be. In addition, digests generally differ depending on the purpose and time limit, but in the conventional cut-and-paste method, an individual digest must be created for each purpose and time limit.

An object of the present invention is to provide a means for dynamically generating a digest as needed in an object-oriented database system storing multimedia data.

[0008]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides storage means for storing multimedia data and information necessary for searching and displaying the multimedia data, and information stored in the storage means. Management means for managing the data, a search means for searching for a necessary portion of the multimedia data using the information stored in the storage means, and digest data generated from a set of the partial data searched by the search means. And an execution unit for displaying on an output device.

The storage means includes a keyword object having keyword information, a target object having information of multimedia data to be searched, and a select object having relation information between the keyword object and the target object. Searches using these objects.

[0010] The target objects include a content object representing the entire content information of the multimedia data and a view object representing the information of the partial data. The search means stores the view object in an element corresponding to the digest data. As a search target.

The keyword object has a synonym attribute, the synonym is treated as one object, and the retrieval means treats the synonym as the same retrieval key. The keyword object has a parent attribute and a descendant attribute,
The keywords are hierarchically managed, and the search means can search all the sub-sets of the specified keywords.

The select object has a weight attribute, has a different weight for each corresponding target object and each keyword object, and gives the same target object a different weight for each keyword.

[0013] When the management means specifies to set a keyword for the target object,
If a keyword object corresponding to the specified keyword exists, a select object having a relationship between the keyword object and the target object is generated. If a keyword object corresponding to the specified keyword object does not exist, a new keyword object is generated. Then, a select object corresponding to the generated keyword object and target object is generated.

When a file name is specified, the search means searches the storage means for a content object corresponding to the file name, and generates a set of corresponding execution objects.

When a keyword name is specified, the search means searches the storage means for a target object corresponding to the keyword, and generates a set of corresponding execution objects.

When a file name and a keyword name are specified, the search means searches the storage means for a content object corresponding to the file name, and searches for a keyword corresponding to the keyword from a set of the content object and a view object corresponding to the content object. And generate a set of corresponding execution objects.

The execution object contained in the storage means has execution information such as a target object to be executed, an execution start time, and an execution duration and information of a priority. To generate a digest and display it based on the execution information.

The execution means determines the priority of the execution object retrieved by the retrieval means from its weight and other information, and generates an execution object describing the priority.

The execution means acquires a specified number of execution object sets from the target execution object set according to the priority described in the execution object.

The execution means, when the target object corresponding to the target execution object is a view object, the content object of the view object is the same, and the data of the portion corresponding to the content object is If they are continuous or overlapping, they are grouped into one execution object as a continuous object.

The execution means acquires a set of execution objects that can be executed for as long as possible without exceeding a designated time according to the priority described in the execution object from the set of target execution objects. I do.

The execution means generates a set of execution objects for generating digest data for a specified keyword within a specified time for a target object corresponding to a target execution object.

[0023] The execution means generates digest data that fits within a specified time for all corresponding view objects when no keyword is specified for a target object corresponding to the target execution object. Generate a set of execution objects to perform

The execution means generates a set of execution objects so as to simultaneously execute a plurality of digest data when specified in the generation of the digest data.

[0025] In the generation of a plurality of digest data, the execution means may generate a set of execution objects so as to simultaneously execute the digest data for each content when specified.

[0026] The execution means converts the multimedia data corresponding to the corresponding target object from a set of execution objects generated for creating a digest into an execution start time and an execution duration described in the execution object. Is displayed continuously according to.

A program for realizing each of the above means by a computer can be stored in a suitable storage medium such as a computer-readable portable medium memory, a semiconductor memory, and a hard disk.

[0028]

FIG. 1 is a diagram showing a configuration example of a system for realizing the present invention. In the example of FIG. 1, the keyword object storage unit 11 and the select object storage unit 1 serve as storage means for storing multimedia data and information necessary for searching and displaying the multimedia data.
2, a target object storage unit 13, an execution object storage unit 14, and a multimedia data storage unit 15.

As a management means for managing information stored in these storage means, an object management unit 21 is provided.
There is a search unit 22 as a search means for searching for a necessary part of the multimedia data using these pieces of information.

In response to an instruction from the input device 25,
There are a digest data generation unit 23 and a data display unit 24 as execution means for generating digest data from a set of partial data searched by the search means and displaying the digest data on the display device 26.

The keyword object storage unit 11 stores a keyword object. The select object storage unit 12 stores a select object. The target object storage unit 13 stores a target object and the like. Execution object storage unit 14
Stores the execution object. The multimedia data storage unit 15 stores multimedia data.

The object management section 21 is means for managing a keyword object, a select object, and a target object. The search unit 22 is a means for searching for a corresponding select object according to a specified file name or keyword and generating a corresponding execution object.

The digest data generating unit 23 is means for generating a set of execution objects as data for digest generation from the set of execution objects generated by the search unit 22.

The data display section 24 is means for displaying the corresponding multimedia data on the display device 26 from the generated set of execution objects. Objects and data handled by this system are keyword objects, target objects, select objects,
It consists of an object for execution and multimedia data.

The keyword object is an object for describing a data search keyword. The target object is an object that describes a search target, and includes a content object and a view object.

The content object is an object related to the content attribute of the multimedia data, and the information includes a file name of the corresponding multimedia data, a data size, a comment of the content, and the like.

The view object includes a feature (for example, a keyword) of partial data as a result of cutting out a part of a series of data from the viewpoint of time and / or space, and data for specifying the partial data (for example, text data. , Data from the beginning of the file to what byte). Therefore, this view object is generated corresponding to the keyword for each partial data.

The select object is an object that associates a keyword object with a target object. The execution object is an object that describes execution information used when the presentation is executed, and is generated from a script file that describes a presentation scenario prepared in advance.

For further details of the content objects used in the present system, see "Multimedia Data Retrieval System, Multimedia Data Object Generating Device and Multimedia Data Management Method" (Japanese Patent Laid-Open No. 9-81443) and " Multimedia data search system, multimedia data scheduling device, multimedia data reproduction method, and reproduction program storage medium ”(JP-A-9-81444)
Are shown.

For details of the view object, see "Multimedia Data Retrieval System" (Japanese Unexamined Patent Application Publication No.
No. 16457). FIG. 2 is a diagram showing an example of each object used in the present system and its attributes and methods.

The keyword object 31 shown in FIG. 2A has an attribute "synonyms", and manages synonyms as one object. The keyword object 31 forms a hierarchical structure of a parent-child relationship, and the identifier of the parent object is described in the parents attribute, and the identifiers of all the lower-level keyword objects are des.
Described in cendants attribute. If there is no parent keyword object or lower keyword object, EMP
TY is described. Also, a plurality of parent keyword objects can be provided. By using the descendants attribute, it is easy to search all the subsets at the time of search.

As a method of the keyword object 31, addParen for setting a parent keyword as a keyword
There is t. Target object 3 shown in FIG.
2 corresponds to the entire multimedia data,
Some correspond to multimedia data portions.

A target object corresponding to the entire multimedia data is called a content object.
The target object corresponding to the part of the multimedia data is called a view object. FIG. 3 is a diagram illustrating a relationship between a content object and a view object. The view object will represent a part of the content object.

The contents attribute contains N when the target object 32 is a content object.
If the IL is a view object, the identifier of the corresponding content object is described. In the views attribute, an identifier of a view object corresponding to the target object 32 is described. generatedTim
The time at which the target object 32 was generated is described in the e attribute. 0 if the target object is a content object in the skipLength attribute
Is a view object, the time from the beginning of the corresponding content to the beginning of the view part is described (see FIG. 3). The timeLength attribute describes the reproduction time of the target object 32. In the fileName attribute, the corresponding multimedia data file name is described when the target object 32 is a content object, and "" is described when the target object 32 is a view object.

The methods include newContents for generating a new content object, setView for setting a view object, addKeyword for setting a keyword and a weight for an object, and digest for generating a digest within a specified time.

The select object 3 shown in FIG.
Reference numeral 3 has a target attribute for describing the target object 32 and a keyword attribute for describing the keyword object 31, and the weight of the target object 32 for the keyword is described in the weight attribute. In the case of a select object corresponding to the target object 32 for which no keyword is set, NIL is described in the keyword attribute.

As a method, viewSe for retrieving a select object matching a specified condition from a database
There is an arch. Execution object 34 shown in FIG.
Indicates the target object 32 to be executed in the target attribute, the execution start time in the startTime attribute, and the timeLeng
The execution duration is described in the th attribute, the weight for determining the execution priority is described in the weight attribute, and the execution priority is described in the priority attribute.

As methods, setPriority for setting a priority order for a set of target execution objects, and selecting a specified number of execution objects in order of priority bes
t, timeLimit to select an execution object in consideration of priority so that it fits within a specified time, processing to combine multiple views corresponding to the same content into one continuous execution object when they overlap TimeOr, parallel, serial, etc. that indicate how to execute the execution object (simultaneously, consecutively, etc.)
And there is do for executing the object for execution.

The details of the processing performed by each of the above methods will be described later with reference to flowcharts. FIG. 4 is a diagram showing a reference relationship between an object and data. The select object 33 refers to the keyword object 31 and the target object 32. The target object 32 refers to the multimedia data 35. The execution object 34 refers to the target object 32.

The flow of the process of creating and executing the digest data in the present invention is as follows. STEP 1: Preparation 1-1: Creation of content object 1-2: Setting of view object 1-3: Setting of keyword and weight 1-4: Structuring of keyword STEP 2: Search STEP 3: Execution 3-1: Setting of priority 3-2: Generation of digest data 3-3: Execution Hereinafter, each will be described in detail.

[STEP 1] Preparation [1-1] Generation of content object (newContent
s) Generate a content object corresponding to the multimedia data.

FIG. 5 is a flowchart of the process for generating a content object. The algorithm will be described with reference to FIG. Target object storage unit 13
A search is made for an object whose fileName attribute value is the file name of the target multimedia data from the set of target objects stored in the database (S10).
1).

The existence of the search result is checked (S102). If the search result exists, the processing is terminated. If the search result does not exist, contents = NIL, views = EMPTY, generatedTime = current time, skipLength = 0, timeLength = total execution time of the target multimedia data, fileName = the target object which is the file name of the target multimedia data, newly created as c,
It is stored in the database of the target object storage unit 13 (S103).

[1-2] Setting View Object (se
tView) Set the view object representing that part in the target object. FIG. 6 is a processing flowchart of setting a view object. The algorithm will be described with reference to FIG.

It is checked whether or not the contents attribute value of the target target object is NIL (S111). If the value is NIL, that is, if the target target object is a content object, the identifier is set to c (S111).
112), if it is not NIL, that is, if it is a view object, its contents attribute value is set to c (S113).

Next, let t be the time from the head of the target target object to the head of the view object, and sl = the skipLength attribute value of the target target object +
Let t, tl = the execution time length of the view object (S114).

A set of objects corresponding to the views attribute value of the object corresponding to c is set to ts (S115).
It is checked whether there is an object whose skipLength attribute value is sl and the timeLength attribute value is tl of the corresponding object among the ts elements (S116), and there is an object whose skipLength attribute value is sl and the timeLength attribute value is tl. If so, the process ends.

If there is no object whose skipLength attribute value is sl and the timeLength attribute value is tl, contents = c, views = EMPTY, generatedTime = current time, skipLength = sl, timeLength = tl, fileName = “” Is newly generated and stored in the database of the target object storage unit 13 (S
117), the identifier of the generated object is added to the views attribute value of the target target object and the object corresponding to c (S118).

[1-3] Setting of Keyword and Weight (addK
eyword) Set keywords and weights for the target object. FIG. 7 is a processing flowchart for setting a keyword and a weight. The algorithm will be described with reference to FIG.

It is checked whether or not a keyword has been designated (S121). If a keyword has been designated, syno is obtained from the set of keyword objects stored in the database of the keyword object storage unit 11.
An object including the specified keyword in the nyms attribute value is searched (S122). If the keyword is not specified, it is determined that k = NIL (S123), and the process proceeds to S127.

It is checked whether or not the search result by the process of S122 exists (S124). If the search result does not exist, a keyword object of synonyms = {specified keyword}, parents = EMPTY, descendants = EMPTY is newly generated. And k
The data is stored in the database of the keyword object storage unit 11 (S125), and the process proceeds to S130.

If there is a search result by the process of S122, it is set as k (S126), and the process proceeds to S127. In the process of S127, an object whose target attribute value is the identifier of the target target object and whose keyword attribute value is k is searched from the set of select objects stored in the database of the select object storage unit 12.

It is checked whether or not the search result exists (S12).
8) If there is a search result, set it as s (S129) and proceed to the process of S131. If there is no search result, a new select object is created by the processing of S130, where targrt = identifier of the target object, keyword = k, weight = default value, the identifier is set to s, and the select object storage unit 12 (S130).

Next, it is checked whether or not the weight is specified (S131).
The value of the weight attribute is set as the weight (S132). [1-4] Structure of keyword (addParent) A keyword object B is set as a parent object of the keyword object A.

FIG. 8 is a flowchart of a keyword structuring process. The algorithm will be described with reference to FIG. First, an identifier of a = A and an identifier of b = B are set (S141).

B is added to the parents attribute value of A (S14)
2) Let ks be the element of the descendants attribute of A (S14)
3) Add a to ks (S144). Also, ks2 =
{B} (S145).

It is checked whether the number of elements of ks2 is 0 (S1
46), if the number of elements of ks2 is not 0, one element k is extracted from ks2, the rest is set as ks2 (S147), and ks is added to the descendants attribute value of the object corresponding to k (S148), and k is added to k. Parents of corresponding object
An attribute element is added to ks2 (S149).

Until the number of elements of ks2 becomes zero, S147-
The process of S149 is repeated, and the process ends when the number of elements of ks2 becomes 0. [STEP 2] Search (viewSearch) There are three methods for searching multimedia data: Search by file name specification Search by keyword specification Search by file name and keyword specification. Returns a set of execution objects as search results.

FIG. 9 is a flowchart of a multimedia data search process. The algorithm will be described with reference to FIG. According to the specified search method (S15
1), a search by specifying a file name (S152), a search by specifying a keyword (S153), or a search by specifying a file name and a keyword (S154).

Depending on whether or not the search result exists (S1
55) If there is a search result, an execution object generation process is performed for the search result (S156), and the result is returned (S157). If there is no search result, EMPTY is returned (S158), and the process ends.

[2-1] Search by File Name In S152 of FIG. 9, by specifying the file name of the multimedia data, the content object (target object) corresponding to the file name is acquired, and the corresponding select object is obtained. Is returned.

FIG. 10 is a flowchart of a search process based on a file name. The algorithm will be described with reference to FIG. From the set of target objects stored in the database of the target object storage unit 13, an object whose fileName attribute value is the specified file name is searched (S161). It is checked whether or not the search result exists (S162). If the search result does exist, a process of acquiring the select object is performed on the search result (S163), and the result is returned (S164). If there is no search result, EMP
TY is returned (S165), and the process ends.

[2-2] Acquisition of Select Object In the acquisition of the select object performed in the process of S163, the corresponding select object is returned to the target target object. If there is a plurality of corresponding select objects, the one with the largest weight attribute value is returned. If there is no corresponding select object, a new select object is generated and returned.

FIG. 11 is a flowchart of a process for acquiring a select object. The algorithm will be described with reference to FIG. A set of identifiers of the target target object is set to ts (S171), and ss = EMPTY is set (S172).

It is checked whether the number of elements of ts is 0 (S17).
3) If the number of elements of ts is not 0, one element t from ts
Is retrieved and the remainder is set to ts (S174), and an object having a target attribute of t is searched from a set of select objects stored in the database of the select object storage unit 12 (S175).

The search result is checked (S176). If the search result does not exist, a select object having target = t, keyword = NIL, weight = default value is newly generated, its identifier is set to s, and the select object is set to s. It is stored in the database of the storage unit 12 (S177). If there is a search result, select the one with the largest weight attribute value and select
(S178). s is added to ss (S179), and S
The process returns to 173.

Until the number of elements of ts becomes 0, S174 to S
If the number of elements of ts becomes 0, ss is returned (S180), and the process ends. [2-3] Search by Keyword In S153 of FIG. 9, by specifying a keyword, a target object having the keyword is acquired, and a set of corresponding select objects is returned.

FIG. 12 is a flowchart of a search process using a keyword. The algorithm will be described with reference to FIG. A set of keyword objects corresponding to the designated keyword is acquired by searching for a keyword object described later (S181).

It is checked whether the return value is EMPTY (S
182), if the return value is EMPTY, return EMPTY (S185) and end. Return value is EMPTY
If not, the return value is set to ks (S183), and ss = E
MPTY is set (S184).

Next, it is checked whether or not the number of elements of ks is 0 (S186). If the number of elements of ks is not 0, one element k is extracted from ks and the remaining is set as ks (S187), and the select object An object whose keyword attribute value is k is searched from the set of select objects stored in the database of the storage unit 12 (S188). If there is a search result (S189), the search result is added to ss (S190).

S186 to S1 until the number of elements of ks becomes 0
If the number of elements of ks is 0, ss is returned (S191), and the process ends. [2-4] Search for Keyword Object In the search for the keyword object performed in S181 of FIG. 12, a keyword object including the specified keyword in the synonyms attribute value is searched from the database.

Using the structure created in the process described in "[1-4] Structuralization of Keyword", keywords (descendant keywords) meaningfully included in the designated keyword are converted to syno.
You can also search for keyword objects included in the nyms attribute. For example, when the keyword “human” is designated, a keyword object with a keyword of “female” or “male” or a personal name can be searched.

FIG. 13 is a flowchart of a keyword object search process. The algorithm will be described with reference to FIG. From the set of keyword objects stored in the database of the keyword object storage 11, an object whose specified synonyms attribute value includes the specified keyword is searched (S201). The search result is checked (S202). If the search result does not exist, E
The MPTY is returned (S204), and the process ends. If there is a search result, the search result is set to ks (S203),
It is checked whether or not the search results refer to descendant keywords (S205).

When referring to descendant keywords, ks2 =
It is set as ks (S206), and it is checked whether the number of elements of ks2 is 0 (S207). If the number of elements of ks2 is not 0, one element k is extracted from ks2, and the rest is set as ks2 (S20).
8) The descendants attribute value of the object corresponding to k is added to ks (S209). The processes of S207 to S209 are repeated until the number of elements of ks2 becomes 0. If the number of elements of ks2 is 0, or if it is not specified to refer to the descendant keyword in the processing of S205, ks is returned (S210), and the process ends.

[2-5] Search by File Name and Keyword In S154 of FIG. 9, by specifying the file name of the multimedia data, a content object (target object) corresponding to the file name is obtained, and Get the view object corresponding to the object. Selects the object corresponding to the specified keyword from the acquired view object and content object, and returns the corresponding select object.

FIG. 14 is a flowchart of a search process using a file name and a keyword. The algorithm will be described with reference to FIG. From the set of target objects stored in the database of the target object storage unit 13, an object whose fileName attribute value is the specified file name is searched (S211).
It is checked whether or not the search result exists (S212). If the search result exists, the search result is set to t (S21).
3) The view attribute value of the object corresponding to t is set to ts, and t is added to ts (S214). A process of acquiring a select object when a keyword described later is specified for ts is performed (S215), and the result is returned (S21).
6).

If there is no search result, the EMPT
Y is returned (S217), and the process ends. [2-6] Acquisition of Select Object When Keyword is Specified Among the target objects, an object having the specified keyword is selected, and the corresponding select object is returned.

FIG. 15 is a flowchart of a process for obtaining a select object when a keyword is specified. The algorithm will be described with reference to FIG. For the specified keyword, the processing described in “[2-4] Search for Keyword Object” is performed (S221).

It is checked whether the return value is EMPTY (S2).
22), if the return value is EMPTY, return EMPTY (S225) and end. Return value is EMPTY
If not, the return value is set to ks (S223), and ss = E
MPTY is set (S224).

Next, for ks, it is checked whether the number of elements of ks is 0 (S226). When the number of elements of ks becomes 0, ss
Is returned (S229), and the process ends. If the number of elements of ks is not 0, one element k is extracted from ks, the rest is set as ks (S227), and a set of identifiers of the target target objects is set as ts (S228). Next, for ts, ts
It is checked whether the number of elements of ts is 0 (S230). If the number of elements of ts is not 0, one element t is extracted from ts,
The remainder is set to ts (S231), and an object whose target attribute is t and keyword attribute is k is searched from a set of select objects stored in the database of the select object storage unit 12 (S232). It is checked whether or not the search result exists (S233). If the search result exists, the search result is added to ss (S234) and S230.
Return to the processing of. If there is no search result, the process returns to S230.

S231 to S2 until the number of elements of ts becomes 0
34 is repeated, and when the number of elements of ts becomes 0, S2
It returns to the process of 26. [2-7] Generation of Execution Object A corresponding execution object is generated from the target select object and returned.

FIG. 16 is a flowchart showing the process of generating an execution object. The algorithm will be described with reference to FIG. The set of identifiers of the target select object is set to ss (S241), and es = EMPTY is set (S242).

It is checked whether the number of elements of ss is 0 (S24).
3) If the number of elements of ss is not 0, one element s from ss
And the remainder is set to ss (S244). A new execution object with target = s.target, startTime = 0, timeLength = s.target.timeLength, weight = s.weight, priority = 0 is newly generated, and es is set to es. It is added (S245).

S244 to S2 until the number of elements of ss becomes 0
When the number of elements of ss becomes 0, es
Is returned (S246), and the process ends. [STEP 3] Execution [3-1] Setting of Priority (setPriority) In generating digest data, a priority is determined for a set of target execution objects. The priority order is determined in descending order of the weight value described in the weight attribute of the execution object. If the weight values are the same, the priority is determined using other information such as the generation time of the view object. Here, an example is shown in which when the weight value is the same value, the priority is set higher in the order of the generation time of the view object (latest order).

FIG. 17 is a processing flowchart for setting the priority. The algorithm will be described with reference to FIG. A set of identifiers of the target execution objects is set to es (S251), and p = 1 (S252).

It is checked whether the number of elements of es is 0 (S25).
3) If the number of elements of es is not 0, the elements of es are sorted in descending order of the value described in the weight attribute (S254),
The first element is extracted from es and set as e, and the rest is set as es (S255).

Of the elements of es, those whose weight attribute value matches the weight attribute value of the object corresponding to e
es2 and the rest are es (S256). e is added to es2 (S257), and the elements of es2 are sorted in descending order of the value described in the generatedTime attribute of the object corresponding to the identifier described in the target attribute (S258).

It is checked whether the number of elements of es2 is 0 (S2
59), if the number of elements of es2 is not 0, the leading element e is extracted from es2, e.priority = p, and the rest is es2
(S260), p = p + 1 (S261), S
Return to 259. S260- until the number of elements of es2 becomes 0
The process of S261 is repeated, and when the number of elements of es2 becomes 0, the process returns to S253.

If it is determined in step S253 that the number of elements of es has become 0, the process ends. [3-2] Digest data generation There are two methods for generating digest data:-Designation of number of views-Designation of playback time. Furthermore, processing to obtain these digest data includes:-processing when view objects overlap during composition-selection of view objects that fit within a specified time. Hereinafter, each of them will be described in detail.

Further, as an extension of the generation of the digest data specifying the reproduction time, a process for simultaneously executing a plurality of pieces of digest data and a process for simultaneously executing the same for each content will be described.

[3-2-1] Generation of Digest Data by Specifying the Number of Views (best) A specified number of execution objects are selected from a set of execution objects according to the priority and returned.

FIG. 18 is a flowchart of a process for generating digest data by specifying the number of views. FIG.
8, the algorithm will be described. A set of the identifiers of the target execution objects is set to es (S271).

It is checked whether or not the priority is set to es (S272). If the priority is not set to es, it is described in “[3-1] Setting of priority (setPriority)”. The processing is applied (S273).

The es are sorted in ascending order of the value described in the priority attribute (S274), and es2 = EMPTY is set (S274).
275), and count = 1 (S276). It is checked whether the number of elements of es is 0 (S277). If the number of elements of es is not 0, it is further checked whether count> specified number (S278). If count> not the specified number,
The first element e is extracted from es, and the rest is set as es (S27
9), add e to es2 (S280), and count = count
As +1 (S281), the process returns to S277.

If the number of elements of es is 0 or count> specified number (S277, S278), es2 is returned (S27).
282), and the process ends. [3-2-2] Generation of Digest Data by Specifying Playback Time (Digest) Digest data that fits within the specified time is generated from the target object, and the resulting set of execution objects is returned. Keywords can also be specified.

FIG. 19 is a flowchart of a process for generating digest data by designating a reproduction time. The algorithm will be described with reference to FIG. The set sum of the views attribute values of the target object corresponding to the target execution object is set to ts (S291). ts has 0 elements
Is checked (S292), and the number of elements of ts is 0
If not, it is checked whether a keyword is specified for ts (S293).

To specify a keyword, refer to “[2-6]
The processing described in "Acquisition of Select Object when Keyword is Specified" is performed (S294), and when the keyword is not specified, the processing described in "[2-2] Acquisition of Select Object" is performed (S295). The return value is checked (S296), and if the return value is not EMPTY, the process described in "[2-7] Creation of an execution object" is performed on the return value, the return value is set to es (S297), and On the other hand, the process described in “[3-1] Setting priority (setPriority)” is applied, and the process of “selection of view object within specified time (timeLimit)” described below is applied to es. es (result) is returned (S300).

If ts does not exist in the process of S292, or the return value is EMPT in the process of S296.
In the case of Y, EMPTY is returned (S301), and the process ends.

[3-2-3] Selection of View Objects Fitting Within Specified Time (timeLimit) Corresponds to the set of execution objects for which priorities have been set in the process of [3-1] Setting Priority. From the set of target objects, select objects within the specified time and return the corresponding execution object as a result. For objects of the same content, processing taking into account time overlap (described later)
To enable continuous playback.

FIGS. 20 and 21 are flowcharts of a process for selecting a view object that fits within the designated time.
The algorithm will be described with reference to FIGS. The set of identifiers of the target execution objects is set to es (S311), es is sorted in ascending order of the value described in the priority attribute (S312), and es2 = EMPTY (S313).

It is checked whether the number of elements of es is 0 (S
314) If the number of elements of es becomes 0, the process proceeds to S320. If the number of elements of es is not 0, the first element e from es
Is extracted, and the rest is set as es (S315). Of the elements of es, the content of the object corresponding to the target attribute value is obtained.
The one whose ts attribute value matches e.target.contents is es3, and the rest is es4 (S316). Furthermore, e and es3
, A process (timeOr) in which view objects overlap, which will be described later, is applied, and the result is set to es5 (S31).
7).

The sum of timeLength of each element of es4 + es5
It is checked whether or not the total value of the timeLength of each element of es4 is larger than the specified time length (S318), and the timeLe of each element of es4 is
If the total value of timeLength of each element of ngth + es5 is not larger than the specified time length, the sum of the elements of es4 and es5 is set to es2 (S319), and the process returns to S314, and es
Total value of timeLength of each element of 4 + time of each element of es5
If the total value of Length is greater than the specified time length, S3
Proceed to step 20.

In the process of S320, es = EMPTY; s
It is assumed that tartTime = 0. It is determined whether or not the number of elements of es2 is 0 (S321).
One element e is taken out from 2 and the rest is set as es2 (S3
22) Among the elements of es2, the contents attribute value of the object corresponding to the target attribute value is e.target.contents
Es3 is the combination of e and e, and the rest is es
2 (S323). Further, es3 is sorted in ascending order of the skipLength attribute value of the object described in the target attribute (S324).

Next, it is checked whether the number of elements of es3 is 0 (S325). If the number of elements of es3 is not 0, es3
Then, the first element e is taken out from e, and e.startTime = e.startTime + startTime is set (S326), and further, startTime = startTime + e.timeLength is set (S327).

E is added to es (S328), and S325 is added.
Return to the processing of. In the process of S325, the processes of S326 to S328 are repeated until the number of elements of es3 becomes 0,
When the number of elements of es3 becomes 0, the process returns to S321.

In the process of S321, the processes of S322 to S328 are repeated until the number of elements of es2 becomes 0,
When the number of elements of 2 becomes 0, es is returned (S32
9), end.

[3-2-4] Processing when View Objects Overlap During Synthesis (timeOr) In a set of execution objects corresponding to the same content, if the corresponding view objects overlap when combining them, the overlapping portion Is regarded as one continuous object. For example, as in the example of FIG.
The corresponding part of the view object is 10-20 seconds,
The set of execution objects that are 40-50 seconds long and the corresponding part of the view object is 35-45se
When an execution object such as c is synthesized, the corresponding part of the view object is 10-20 seconds, 35
A set of execution objects is created that is -50 seconds.

FIG. 23 is a processing flowchart in the case where view objects overlap at the time of composition. The algorithm will be described with reference to FIG. The set sum of the identifiers of the execution object to be combined and the execution object to be combined is es (S331), and es is sorted in ascending order of the skipLength attribute value of the object corresponding to the target attribute (S332). es2 = EMPTY (S333),
The first element e is extracted from es, and the remaining element is set as es (S3
34).

It is checked whether or not the number of elements of es is 0 (S335). If the number of elements of es is not 0, the first element e2 is extracted from es and the rest is set as es (S336). e.
It is checked whether startTime + e.timeLength <e2.startTime (S327), and e.startTime + e.timeL
If ength <e2.startTime is not satisfied, the process proceeds to S338.

In the process of S338, e.startTime + et
Check if imeLength <e2.startTime + e2.timeLength, and check e.startTime + e.timeLength <e2.st
If artTime + e2.timeLength, e.timeLength = e2.startTime + e2.timeLength-e.
It is set as startTime (S339). e.startTime + e.timeLength <
If not e2.startTime + e2.timeLength, S3
It returns to the process of 35.

On the other hand, in the process of S337, e.startTime +
If e.timeLength <e2.startTime, e is es2
And e = e2 (S340), and the process returns to S335.

In the process of S335, the number of elements of es is 0
The processing of S335 to S339 is repeated until
When the number of elements becomes zero, e is added to es2 (S34).
1) Return es2 (S342) and end.

[3-2-5] Generation of Digest Data for Simultaneous Execution of a Plurality of Digests Digest data that fits in a specified number of times or less within a specified time is generated from a target object.
Return the result set of execution objects. This algorithm performs the following processing by changing a part of the algorithm of the processing of “[3-2-2] Generation of digest data by designating reproduction time”.

FIGS. 24 and 25 are processing flowcharts for generating digest data for simultaneously executing a plurality of digests. The algorithm will be described with reference to FIGS.

The processing of S351 to S358 and the processing of S359 shown in FIG. 24 correspond to S291 to S291 of the processing flow of digest data generation by designating the reproduction time shown in FIG.
Since the processing is the same as the processing in S298 and the processing in S301, the description is omitted.

In FIG. 25, after proceeding to the processing of S358, v = specified number and es2 = EMPTY (S3
60). It is checked whether v = 0 (S361). If v = 0, it is checked whether the number of elements of es is 0 (S362). If the number of elements of es is not 0, the processing of “selection of view object that fits within specified time described in [3-2-3] (timeLimit)” is applied to es (S363), and es ( (Result) is added to es2 (S36)
4). Further, of es, the execution object corresponding to the view object to which timeLimit is applied in S363 is removed, and the rest is set to es (S365). v = v-1
(S366), and returns to S361.

In the processing of S361 and S362, v =
When 0 or the number of elements of es becomes 0, return es2 (S
367), end. [3-2-6] Generation of Digest Data for Simultaneously Performing Digests for Each Content From the target object, generate digest data that fits within a specified time for each specified number of contents or less and execute the result Returns a set of objects for use. This algorithm performs the following processing by changing a part of the algorithm of the processing of “[3-2-2] Generation of digest data by designating reproduction time”.

FIGS. 26 and 27 are flowcharts of processing for generating digest data for simultaneously executing a plurality of digests. The algorithm will be described with reference to FIGS.

The processing of S371 to S378 and the processing of S379 shown in FIG. 26 correspond to the processing flow of S291 to S digest of the digest data generation by designating the reproduction time shown in FIG.
Since the processing is the same as the processing in S298 and the processing in S301, the description is omitted.

In FIG. 27, after proceeding to the processing of S378, v = specified number, es2 = EMPTY (S3
80). It is checked whether v = 0 (S381), and if not v = 0, it is checked whether the number of elements of es is 0 (S382). If the number of elements of es is not 0, the elements of es are extracted and set to e, and the rest are set to es (S383).

Next, of the elements of es, the contents attribute value of the object corresponding to the target attribute value is e.target.
Let es3 be the one that matches contents and es the rest
e is added to es3 (S384). For es3, "[3
-2-3] Applying the process of “selection of view object that fits within specified time (timeLimit)” (S385), es3
(Result) is added to es2 (S386).

Assuming that v = v−1 (S387), S381
Return to In the judgments of S381 and S382, if v = 0 or the number of elements of es becomes 0, es2 is returned (S38
8), end.

[3-3] Execution (do) An execution instruction is issued to a set of execution objects. FIG. 28 is a processing flowchart of the execution. The algorithm will be described with reference to FIG.

A set of object execution object identifiers is
es (S391), and fork & ex for each element e of es
ec (create / execute child process) (S392-S3
96).

That is, fileName = e.target.fileName skipLength = e.target.skipLength startTime = e.startTime timeLength = e.timeLength (S393-S395), and the player is set to startT.
ime is started, and timeLength is reproduced from the position skipped by skipLength of the multimedia data corresponding to fileName (S396 to S398).

The player is means for reproducing the multimedia data of the corresponding fileName and displaying it on the display device. The details of the player are described in, for example, "Multimedia Data Search System, Multimedia Data Scheduling Apparatus, Multimedia Data Reproduction Method, and Reproduction Program Storage Medium" (JP-A-9-814).
No. 44).

[0137]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Two examples will be given as embodiments of the present invention. [Example 1] Reproduction of digest data specifying the number of views Fig. 29 shows an example of a script for reproducing the digest data specifying the number of views.

Here, a view object having a designated keyword ("software") is searched from the database of the target object storage unit 13, and the search results are reproduced six by six in descending order of priority. The meaning of each line in FIG. 29 is as follows.

1: Define a set of objects os, os2, os3. 2: Set os to empty (EMPTY). 3: Retrieve the view object of the video whose keyword is "software" from the database and return the result as os2
And

4: The following processing is repeated until os2 becomes empty. 5: Six objects are selected in descending order of priority from os2, and they are set to be executed in parallel (simultaneously start execution), and the result is set to os3.

6: os3 is set to be executed continuously and added to os. 7: Remove os3 from os2. The process returns to 8: 4.

9: os is executed. [Embodiment 2] Reproduction of digest data specifying reproduction time FIG. 30 shows an example of a script for reproducing the digest data specifying the reproduction time.

Here, the content object corresponding to the multimedia data of the specified file name ("Introduction to Laboratory") is searched from the database of the target object storage unit 13, and the digest specifying the time and the keyword is specified in the search result. Generate and play. FIG.
The meaning of each line is as follows.

1: Define a set of objects os. 2: Search the database for a video object (content object) whose file name is "Research Introduction" and set the result to os.

3: From the view object of os, in which the keyword is "software", the objects are selected such that the entire object is within 5 minutes (300 seconds) in descending order of priority, and the result is os.

4: Execute os.

[0147]

As described above, according to the present invention, a necessary portion of multimedia data stored in a database is taken out and reproduced in a required number according to the priority order, or collected within a required time. Or play it as a digest.

Since this digest is dynamically generated, even if it is desired to generate the digest by changing the reproduction time or the keyword depending on the purpose, the digest is automatically generated easily, which is much larger than the conventional digest data created by cutting and pasting. You can save time and effort.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a system configuration for realizing the present invention.

FIG. 2 is a diagram illustrating an example of an attribute and a method of an object of each class used in the present system.

FIG. 3 is a diagram illustrating a relationship between a content object and a view object.

FIG. 4 is a diagram showing a reference relationship between an object and data.

FIG. 5 is a processing flowchart of generating a content object.

FIG. 6 is a processing flowchart of setting a view object.

FIG. 7 is a processing flowchart of setting a keyword and a weight.

FIG. 8 is a processing flowchart of keyword structuring.

FIG. 9 is a flowchart of a search process.

FIG. 10 is a flowchart of a search process based on a file name.

FIG. 11 is a processing flowchart of acquiring a select object.

FIG. 12 is a flowchart of a search process using a keyword.

FIG. 13 is a flowchart of a keyword object search process.

FIG. 14 is a processing flowchart of a search using a file name and a keyword.

FIG. 15 is a processing flowchart of acquiring a select object when a keyword is specified.

FIG. 16 is a processing flowchart of generating an execution object.

FIG. 17 is a processing flowchart of setting a priority.

FIG. 18 is a processing flowchart for generating digest data by designating the number of views.

FIG. 19 is a processing flowchart of generation of digest data by designating a reproduction time.

FIG. 20 is a flowchart (1) of a process of selecting a view object that fits within a designated time.

FIG. 21 is a flowchart (2) of a process of selecting a view object that fits within a designated time.

FIG. 22 is a diagram illustrating an example of a case where view objects overlap during synthesis.

FIG. 23 is a flowchart of a process in a case where view objects overlap during synthesis.

FIG. 24 is a flowchart (1) of a process of generating digest data for simultaneously executing a plurality of digests.

FIG. 25 is a flowchart (2) of a process of generating digest data for simultaneously executing a plurality of digests.

FIG. 26 is a flowchart (1) of a process of generating digest data for simultaneously executing a digest for each content.

FIG. 27 is a flowchart (2) of a process of generating digest data for simultaneously executing a digest for each content.

FIG. 28 is a processing flowchart of execution.

FIG. 29 is a diagram illustrating an example of a script for reproducing digest data in which the number of views is specified by using a method best.

FIG. 30 is a diagram illustrating an example of a script for reproducing digest data with a specified reproduction time using a method digest.

[Explanation of symbols]

 11 Keyword Object Storage Unit 12 Selected Object Storage Unit 13 Target Object Storage Unit 14 Execution Object Storage Unit 15 Multimedia Data Storage Unit 21 Object Management Unit 22 Search Unit 23 Digest Data Generation Unit 24 Data Display Unit 25 Input Device 26 Display Device 31 Keyword object 32 Target object 33 Select object 34 Execution object 35 Multimedia data

Claims (7)

[Claims] 1. An information processing apparatus for retrieving and outputting multimedia data from a database, comprising: storage means for storing multimedia data and information necessary for searching and displaying the multimedia data; Management means for managing information to be retrieved, search means for searching for a required portion of multimedia data using information stored in the storage means, and digest data from a set of partial data searched by the search means. And an execution means for generating the data and outputting the generated data to an output device. 2. The information stored in the storage means includes a keyword object having keyword information, a target object having information of multimedia data to be searched, and a select having relation information between the keyword object and the target object. 2. The multimedia data presentation system according to claim 1, further comprising an object, wherein said retrieval means retrieves using said object to generate an execution object. 3. When the management means specifies to set a keyword for the target object, if there is a keyword object corresponding to the specified keyword, a select object having a relationship between the keyword object and the target object is set. And generating a new keyword object if there is no keyword object corresponding to the designated keyword object, and generating a select object corresponding to the generated keyword object and the target object. The described multimedia data presentation system. 4. The search means searches the storage means for a content object corresponding to a file name when a file name is specified, or searches for a target object corresponding to a keyword when a keyword name is specified. When a set of execution objects corresponding to a file name and a keyword name are specified by searching from the storage unit, a content object corresponding to the file name is searched from the storage unit when a file name and a keyword name are specified. 3. The multimedia data presentation system according to claim 2, wherein a set corresponding to the keyword is selected from a set of view objects corresponding to the object, and a set of corresponding execution objects is generated. 5. The execution means, for a target object corresponding to a target execution object,
If a keyword is specified, a set of execution objects for generating digest data within the specified time for the specified keyword is generated. If no keyword is specified, all view objects corresponding to the specified keyword are targeted. 3. The multimedia data presentation system according to claim 2, wherein a set of execution objects for generating digest data that fits within a designated time is generated. 6. The execution means, based on a set of execution objects generated for creating a digest, converts multimedia data corresponding to a corresponding target object into an execution start time and an execution start time described in the execution object. 3. The multimedia data presentation system according to claim 2, wherein the display is continuously performed according to the duration. 7. A storage medium for storing a program used for an information processing apparatus for searching for and outputting multimedia data from a database, wherein the storage medium stores stored multimedia data and information necessary for searching and displaying the multimedia data. A function of managing, a function of searching for a necessary part of multimedia data using the managed information, and a function of generating digest data from a set of the searched partial data and outputting the digest data to an output device. A storage medium for storing a program for realizing a multimedia data presentation system characterized by storing a program for realizing the program on a computer.