KR20090063092A - Metadata based indexing and retrieving apparatus and method - Google Patents

Abstract

A metadata base index/search apparatus and a method thereof for improving search speed and search performance are provided to use a query for the metadata search generated from an inputted user query. A metadata extracting unit(10) extracts metadata from input data. A metadata indexing unit(20) produces index data in which a search object is from the extracted meta data. An index data storage by metadata(30) stores the generated index data according to meta data. A query analyzing unit of the quality produces the query for the metadata search from the query inputted from the user. A metadata searching unit searches index data corresponding to the query for the generated metadata search in the meta data index data storage module.

Description

Metadata-based indexing and retrieval device and its method {METADATA BASED INDEXING AND RETRIEVING APPARATUS AND METHOD}

The present invention relates to the storage and retrieval of data files, and more particularly, to the storage and retrieval of data files in portable terminals.

The present invention is derived from the research conducted as part of the IT growth engine technology development project of the Ministry of Information and Communication and the Ministry of Information and Communication Research and Development. [Task Management Number: 2005-S-117-03, Title: Development of Intelligent Personal Media Managing Technology for Ubiquitous Environment].

Data retrieval in mobile terminals such as mobile phones was mainly limited to phone numbers or address books. However, due to the increase in portable terminal capacity, various search demands have emerged as personal information such as memos, calendars, and text messages, and media files such as images, voices, and videos are also stored and managed in the portable terminal.

Currently, the search of the mobile terminal is focused on web search using a server-client structure and embedded in the mobile device itself, so that a simple DB system or a tree-type simple index is used to search the data of the mobile device. I am using a structure. However, since the search of the mobile terminal has such a simple structure, it cannot index / search complex or large amounts of data.

As described above, while the storage space of the portable terminal increases, there is a limit in the searching ability of the data files stored therein. That is, in the related art, since the portable terminal provides only a search for a simple search target such as a telephone number or a file name, the user's access to personal information or media files is inconvenient in the portable terminal.

An object of the present invention is not to simply store and retrieve a large amount of data stored in the mobile terminal as in the past, but to index and store the data in metadata units meaningful to the data, thereby limiting the limitation of the mobile terminal. The present invention provides a metadata-based indexing and retrieval device and a method thereof that enable efficient data retrieval even in an environment.

In order to achieve the above object, the metadata-based indexing and retrieval apparatus of the present invention includes a metadata extraction unit for extracting metadata from input data input to a mobile terminal, and an index to be searched for the input data. A metadata indexing unit for generating data from the extracted metadata, a metadata indexing unit for storing the generated index data for each metadata, and a query for metadata searching from a query input from a user And a query analyzing unit and a metadata searching unit searching the index data corresponding to the generated metadata searching query in the index data storage unit for each metadata and providing the search result.

The input data may include personal information and media files. The metadata extractor may include a personal information metadata extractor for extracting metadata related to the personal information and a media file metadata extractor for extracting metadata related to the media file. In this case, the personal information may include a schedule, and the metadata about the schedule may include at least one of a title, a classification, a description of the schedule, and last modified date time information. In addition, the media file may include an audio file, and the metadata of the audio file may include information about one or more of a singer, an album name, a genre, a play time, a play frequency, an album production year, lyrics, a copyright, and a format.

The metadata may include common metadata common to input data and unique metadata unique to each input data.

The metadata index unit may generate the index data from the metadata by morphological analysis or n-gram analysis.

The index data storage unit for each metadata may store the index data for each metadata in one place so that input data may be searched without specifying metadata.

The query analyzer may generate different metadata search queries according to the type of the input query. In this case, the type of the input query includes an integrated keyword query including a keyword to be used for a search for all metadata and a metadata specification query including a keyword to be used for a search for specific metadata. Optionally, natural language queries may be further included. The natural language query may be converted into either the integrated keyword query or the metadata designation query.

The metadata search unit may search the index data storage unit for each metadata by a different search method according to a form of the metadata search query. That is, when the metadata search query is a query for all metadata, the metadata search unit searches for index data of all metadata, and the metadata search query searches for specific metadata. In the case of a target query, the metadata search unit searches for the index data of the specific metadata. In addition, the specific metadata may be one or a plurality of metadata.

In addition, in order to achieve the above object, the metadata-based indexing and retrieval method of the present invention is a metadata-based indexing and retrieval method of personal information and media files stored in a mobile terminal, and by means of a metadata extracting unit, And receiving a media file and extracting metadata, and generating, by the metadata indexing unit, index data to be searched at the time of searching for the personal information and the media file from the extracted metadata; Storing, by the index data storage unit for each metadata, the generated index data for each metadata; generating, by the query analyzer, a query for metadata searching from a query input from a user; and searching for metadata By the metadata, the index data corresponding to the generated metadata search query Searching in the index-specific data storage unit for each data and providing the search result.

According to the present invention, the personal information and the media file stored in the mobile terminal are indexed and stored by metadata, respectively, and the metadata search query is generated from the input user query and the search is performed based on the vast amount of portable data. Fast and efficient retrieval of data from the terminal is possible.

In addition, according to the present invention, the search speed and the search performance can be improved by performing indexing and searching using a lot of extractable metadata.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a metadata-based index and search apparatus according to an embodiment of the present invention. The device of this embodiment indexes and retrieves personal information and media files stored in the personal digital assistant based on the metadata. Hereinafter, the present embodiment will be described with respect to personal information or a media file. However, any input data that is input and stored in the portable terminal can be the object of the present embodiment.

The apparatus of this embodiment includes a metadata extractor 10, a metadata indexer 20, an index data storage unit 30 for each metadata, a query analyzer 40, and a metadata searcher 50. do.

The metadata extracting unit 10 extracts common metadata common to personal information or a media file in the personal portable terminal, and extracts unique metadata unique to the personal information or media file. In addition, the metadata extractor 10 includes a personal information metadata extractor for extracting metadata related to personal information, and a media file metadata extractor for extracting metadata related to a media file.

The metadata indexing unit 20 analyzes individual metadata extracted from the metadata extracting unit 10 or individual metadata of the media file by morphological analysis or n-gram analysis method and indexes using an independent storage structure. . In addition, for the integrated search, all metadata index data related to personal information or media files are stored in the metadata index data storage unit 30.

The index data storage unit 30 for each metadata stores index data for each metadata related to personal information or media files provided from the metadata index unit 20 in a certain storage structure in a tree form.

The query analyzing unit 40 generates a query for metadata search according to the form of the query 41 input by the user. The query 41 entered by the user includes an integrated keyword query, a metadata specification query, and a natural language query. An integrated keyword query is a query composed of keywords used in a general search without information on specific metadata. The metadata specification query refers to a query in which specific metadata and keywords are matched. Natural language queries refer to queries composed of natural language in the form of syntax, not keywords of nouns. The natural language query goes through a morphological analysis and becomes an integrated keyword query or metadata specification query. As a result, the metadata retrieval query is generated in the form of either an integrated keyword query or a metadata specification query.

The metadata search unit 50 receives a query for metadata search generated from the query analysis unit 40, searches the index data storage unit 30 for each metadata, finds index data for each metadata, and searches the search result. Provide 51 to the user.

The metadata retrieval unit 50 divides the metadata retrieval query received from the query analysis unit 40 into an integrated query, one metadata specifying query, and two or more metadata specifying queries according to the user's intention. Performs metadata-based searches.

2 is a flowchart illustrating a metadata-based indexing and searching method of the present invention for indexing and storing metadata of personal information and media files stored in a mobile terminal, and receiving a user query and searching the metadata query.

When the personal information or the media file generated in the portable terminal is input to the metadata extractor 10, the metadata extractor 10 extracts individual metadata related to the personal information and the characteristics of the media file (S100).

When the individual metadata related to the personal information and the media file extracted from the metadata extraction unit 10 are input to the metadata indexing unit 20, the metadata indexing unit 20 may perform metametry through morphological analysis or n-gram analysis. Personal information and media files are indexed by data, and index data for each metadata is stored in the index data storage unit 30 for each metadata (S200).

When the mobile terminal receives a user query 41 regarding personal information or a media file to be searched by the user through the query analyzer 40, the query analyzer 40 analyzes the user's query 41 to search for metadata. The query is transmitted to the metadata search unit 50 (S300).

The metadata search unit 50 receives a query for metadata search from the query analysis unit 40, searches the index data storage unit 30 for each metadata, finds index data for each metadata, and retrieves the search result 51. ) Is output to the user (S400).

The present invention is characterized in that the metadata indexing unit 20 stores personal information or index data for each metadata associated with a media file to the storage unit 30, and the personal information desired by the metadata searching unit 50. And retrieve the media file and provide the search results to the user.

3 is a flowchart illustrating a method of extracting metadata from a metadata extracting unit.

The metadata extracting unit 10 receives the personal information or the media file to be extracted (S101), and first extracts common metadata that can be commonly extracted from the information of the mobile terminal (S102).

Examples of common metadata are shown in Table 1 below, and these metadata can be added and deleted according to a search need.

Metadata name Explanation MediaType Media type File Filename Ext File extension Path File absolute path Time Creation or modification time of the file keyword Keyword of file (if entered into file) descrip File description (if entered in file) Title File title (if entered in file) creator User who created the file

Common metadata includes media types, file names, file extensions, file absolute paths, and file creation or modification times, and includes keywords, file details, file titles, and user information that created the file. If it is entered in the file, that information is also included.

After extracting the common metadata, the metadata extraction distributor determines what metadata to extract using the characteristics of the file in order to extract unique metadata unique to each individual information or media file (S103).

The metadata extraction distributor selects an extractor that matches the input data.

If the extraction target is personal information, the personal information metadata extractor extracts metadata related to personal information (S104).

If the extraction target is a media file, the media file metadata extractor extracts metadata related to the media file (S105).

Personal information includes schedules, to dos, notes, contact numbers, text messages, and the like, and can be added and deleted as needed. The media file includes information of one or more of text, image, voice, and video data.

An example of metadata for a schedule is shown in Table 2 below, which includes the title, classification, description of the schedule, and time of last modification date.

Metadata name Explanation Title title Category Classification Note Description of the event ModifiedDate Last Modified Date Time

The media file may include text, images, audio, and video, and may define metadata according to the characteristics of the media.

Table 3 below is an example of metadata about an audio media file.

Field name Explanation artist Singer, artist album Album name genre Genre ptime Playing time freq Refresh rate pyear Album Year lyric lyrics rights Copyright format format

The metadata of the audio media file includes information of one or more of artist, album name, genre, play time, play frequency, year of album production, lyrics, copyright, and format.

4 is a flowchart illustrating a method of indexing for each metadata in the metadata indexing unit.

The metadata indexing unit 20 receives metadata extracted from personal information or a media file (S201) and extracts index units by analyzing metadata by morphological analysis or n-gram analysis (S202).

Index unit extraction in the present invention used two methods, morphological analysis or n-gram analysis.

In the case of using the first morphological analysis, the root of the word can be known. Therefore, even if a variant of the word occurs, such as a survey attachment in the case of Hangul and a change in the plural or tense in the case of English, the root can be found using the root.

Secondly, the n-gram analysis method is an effective method for reducing the cpu occupancy of the mobile terminal when morphological analysis is not feasible.

(Metadata, index unit) analyzed in this way in the metadata indexing unit 20 is indexed for each individual metadata (S203), the index data for each metadata generated in this way index data storage unit 30 for each metadata Save to (S204).

The index data is stored so that indexes for the number of metadata are generated by indexing for each individual metadata. In addition, all indexed metadata is stored in one place so that integrated search is possible without specifying metadata.

Morphological analysis is the analysis of meaningful units of language. In Korean, the morphological analysis is divided into formal morphemes and real morphemes. Extraction of index units for search is usually only for nouns.

The following is an example of Korean morpheme analysis and index unit extraction.

Sentence: Photo taken on a business trip to Chungcheong-do

Stemming results:

[To Chungcheongdo] Chungcheongdo / NR + / JO

[Business trip] Capped / NN + / JO

[Go] Go / VV + Go / EM

[Taken] Snip / VV + Silver / EM

[Photos] Photo / NN

(NN: noun, NR; proper noun, JO: probe, VV: verb, EM: mother)

Index unit:

Chungcheong-do

[Business trip] business trip

[Go]

[hewn]

[Photo] photo

Sentences presented by morphological analysis are indexed by analyzing only the Chungcheong-do, business trip, and photographs in index units.

n-gram analysis is a method of extracting index units using n consecutive syllables and is used when morphological analysis is difficult to use. N-gram analysis has the advantage of being processed with a simple algorithm when it is impossible to develop an expensive stemmer, but extracts useless data in units of indexes.

The following is an example of a bi-gram using two consecutive syllables.

Sentence: Photo taken on a business trip to Chungcheong-do

Bi-gram Analysis Index Unit:

[Chungcheong-do] Chungcheong, Qingdao, Doe

[Business trip] business trip, chapter

[Go] go

Taken

[Photo] photo

The most widely used bi-grams are ' Chungcheong ', 'Qingdao', ' Doe ', ' Business trip ', 'Changing', 'Go', 'Take' and ' Picture ' The drawback is that data that does not need to be indexed except underlined index units or that should not be indexed is extracted in index units.

5 is a flowchart illustrating a method of analyzing a query according to a user's query in the query analyzer.

When the user inputs a query expressing an object to be searched for by the user (S301), the query analyzing unit 40 generates a metadata search query that is actually searched through a morphological analysis process (S302) (S303). ).

The user's query can be of three types: integrated keyword query, metadata specification query, and natural language query. Each final query is as follows.

(1) Integrated Keyword Query: It is used when the user wants to search the entire metadata and consists of keywords used in general search.

Ex) mobile phone

(2) Metadata specification query: When the user specifies to search only in specific metadata, query in the form of 'metadata name: keyword'.

Ex) artist: ivy

(3) Natural language query: User queries in natural language instead of noun keyword, and it becomes integrated keyword query or metadata specification query using morphological analysis process and query rule.

Ex) File extension mp3-> Ext: mp3

6 is a flowchart illustrating a method of searching metadata and providing a search result in the metadata searching unit.

The metadata search unit 50 actually searches by using the metadata search query generated from the query analyzer 40 (S401), and inquires the index data storage unit 30 for each metadata (S402).

The search using the metadata search query S401 classifies the query into three types below to determine which index data 30 to search from among indexed data for each metadata.

* Integrated query: search for integrated index data

* 1 metadata specified query: search for index data of specified metadata

* 2 or more metadata specified queries: search by combining only two or more specified metadata index data

In the same manner as above, the metadata search unit 50 searches only the designated index data and provides the search result 51 to the user of the mobile terminal (S403).

The personal information and media file metadata-based indexing and retrieval device extracts the predefined metadata necessary for retrieving the personal information and the media file stored in the mobile terminal through the above process, and by morphological analysis and n-gram analysis Index and store by individual metadata, analyze the input user query to search the metadata to finally provide the user with the desired search results.

As described above, although described with reference to a preferred embodiment of the present invention, those skilled in the art without departing from the spirit and scope of the present invention described in the claims In the present invention can be carried out by various modifications or variations.

1 is a block diagram of a metadata-based index and retrieval apparatus of personal information and media files stored in a personal portable terminal according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a method of indexing and storing metadata of personal information and media files stored in a mobile terminal according to the present invention, and receiving a user query and searching the metadata query.

3 is a flowchart illustrating a method of extracting metadata from a metadata extracting unit.

4 is a flowchart illustrating a method of indexing for each metadata in the metadata indexing unit.

5 is a flowchart illustrating a method of analyzing a query according to a user's query in the query analyzer.

6 is a flowchart illustrating a method of searching metadata and providing a search result in a metadata searching unit.

Claims (10)

A metadata extraction unit for extracting metadata from input data input to the mobile terminal; A metadata index unit which generates, from the extracted metadata, index data to be searched when searching for the input data; An index data storage unit for each metadata for storing the generated index data for each metadata; A query analyzer for generating a query for metadata search from a query input from a user; And a metadata search unit for searching the index data corresponding to the generated metadata search query in the metadata index data storage unit and providing a search result. The method according to claim 1, The input data includes personal information and a media file, wherein the metadata extracting unit includes a personal information metadata extractor for extracting metadata related to the personal information, and a media file for extracting metadata related to the media file. A metadata-based indexing and retrieval device comprising a metadata extractor. The method according to claim 1, And the metadata includes common metadata common to the input data and unique metadata unique to each input data. The method according to claim 1, The metadata indexing unit generates the index data from the metadata by morphological analysis or n-gram analysis. The method according to claim 1, The metadata-based index data storage unit stores metadata-based index data in one place so that input data can be searched without specifying metadata. The method according to claim 1, The query analyzer generates different metadata search queries according to the type of the input query, and the type of the input query includes a keyword to be used for a search for all metadata. And a metadata specification query that includes a keyword to be used for a search for a specific metadata. The method according to claim 6, The form of the input query further includes a natural language query, and the query analyzing unit converts the natural language query into either the integrated keyword query or the metadata specification query. The method according to claim 1, The metadata retrieval unit is a metadata-based index and retrieval device for searching the index data storage for each metadata by a different retrieval method according to the form of the metadata search query. The method according to claim 8, When the metadata search query is a query for all metadata, the metadata search unit searches for index data of all metadata, and the metadata search query targets specific metadata. In the case of a query, the metadata retrieval unit performs a search on the index data of the specific metadata. A metadata-based indexing and retrieval method of personal information and media files stored in a mobile device, Extracting metadata by receiving the personal information and the media file by a metadata extracting unit; Generating, by the metadata indexing unit, index data to be searched in the search for the personal information and the media file from the extracted metadata; Storing, by the index data storage unit for each metadata, the generated index data for each metadata; Generating, by the query analyzer, a query for metadata search from a query input from a user, And searching, by a metadata search unit, index data corresponding to the generated metadata search query in the metadata index storage unit for each metadata to provide a search result.

Images