JP2010231271A - Content search device, content search method, and content search program - Google Patents

Abstract

A content search apparatus, a content search method, and a content search program that can more easily search for desired content.
A content search apparatus includes a first storage unit storing a plurality of contents and content information including a plurality of items, content information of an arbitrary selected content, and content of other contents An extraction unit that determines whether all or part of the information matches or does not match, and extracts a plurality of contents from the first storage unit, and the extracted content information is extracted by the content information and extraction unit of the selected content. The degree of matching with the content information of the selected content is determined, a degree of association calculation unit that calculates the degree of association with the selected content, and the arrangement position of the extracted content is calculated according to the calculated degree of association An arrangement position calculation unit; and a display unit that displays a plurality of contents extracted by the extraction unit at the calculated arrangement position.
[Selection] Figure 1

Description

  The present invention relates to a content search apparatus, a content search method, and a content search program for searching for desired content from a plurality of contents.

  In the PC, the search function is a general technique widely used in daily work. In the PC, a user searches for a desired object (for example, content) by inputting a keyword. On the other hand, some AV devices such as a TV, a DVD recorder, and an HDD recorder have a content search function (Patent Document 1).

JP 2008-242504 A

In recent years, the number of contents that can be searched for by users has increased rapidly due to an increase in the capacity of storage devices built in or externally attached to AV equipment and the connection of AV equipment to the Internet. For this reason, it is difficult to search for a desired content from this enormous number of contents.
In view of the above, an object of the present invention is to provide a content search device, a content search method, and a content search program that can more easily search for desired content.

  A content search device according to an aspect of the present invention includes a first storage unit in which a plurality of contents and content information including a plurality of items corresponding to the plurality of contents are stored, and an arbitrary one of the plurality of contents An instruction receiving unit that receives a selection instruction to select the content, and the content information of the selected content and the content information of the content stored in the first storage unit are determined to match or not match An extraction unit that extracts a plurality of contents from the first storage unit based on the determination result, and content information of the selected content and content of the content extracted by the extraction unit for each content extracted by the extraction unit Judge whether the item matches information or not, and based on the result of the decision, relevance to the selected content A degree-of-association calculation unit that calculates the arrangement position of the content extracted by the extraction unit according to the degree of association calculated by the calculation unit, and an arrangement position calculated by the arrangement position calculation unit And a display unit for displaying a plurality of contents extracted by the extraction unit.

  A content search method according to an aspect of the present invention selects arbitrary content from a first storage unit in which a plurality of contents and content information including a plurality of items corresponding to the plurality of contents are stored. Determining whether or not all or a part of the content information of the selected content and the content information of the content stored in the first storage unit are coincident or inconsistent, and selecting a plurality of contents based on the determination result The step of extracting from the first storage unit, and for each extracted content, the content information of the selected content and the content information of the content extracted by the extraction unit are determined to match, and the determination result And calculating the relevance level with the selected content based on the Calculating a position of the extracted contents by part, on the calculated position by position calculation unit comprises a step of displaying a plurality of content extracted by the extraction unit.

  A content search program according to an aspect of the present invention includes a computer that includes a first storage unit in which a plurality of contents and content information including a plurality of items corresponding to the plurality of contents are stored; An instruction accepting unit that accepts a selection instruction for selecting arbitrary content from the content, and the content information of the selected content and the content information of the content stored in the first storage unit, all or some of the items match or mismatch And extracting the plurality of contents from the first storage unit based on the determination result, and the content information of the selected content and the extraction unit for each content extracted by the extraction unit It is determined whether the item matches the content information of the content, and the selected item is selected based on the determination result. A degree-of-association calculation unit that calculates a degree of relevance to the content, a position-position calculation unit that calculates a position of the content extracted by the extraction unit according to the degree of relevance calculated by the calculation unit, and a position-position calculation unit The display unit is operated as a display unit that displays a plurality of contents extracted by the extraction unit at the arrangement position calculated by the above.

  According to the present invention, it is possible to provide a content search device, a content search method, and a content search program that can more easily search for desired content.

It is the figure which showed the structure of the content search device which concerns on 1st Embodiment. It is a figure showing functional composition of a contents search device concerning a 1st embodiment. It is the figure which showed the example of the reference | standard content selection screen. It is the figure which showed the other example of the reference | standard content selection screen. It is the figure which showed an example of the search condition input screen. It is the figure which showed an example of the search condition input screen. It is the figure which showed an example of the relationship between the number of contents, and a relevance degree. It is the figure which showed another example of the relationship between the number of contents and a relevance degree. It is the flowchart which showed operation | movement of the content search apparatus which concerns on 1st Embodiment. It is the figure which showed the structure of the content search device which concerns on 2nd Embodiment. It is the figure which showed the function structure of the content search device which concerns on 2nd Embodiment. It is the figure which showed an example of the information memorize | stored in search log | history DB. It is the figure which showed an example of the information memorize | stored in search log | history DB. It is the flowchart which showed operation | movement of the content search apparatus in 2nd Embodiment. It is the figure which showed the function structure of the content search device which concerns on 3rd Embodiment. It is the figure which showed operation | movement of the content search apparatus which concerns on 3rd Embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(First embodiment)
First, an outline of the content search apparatus 1 according to the first embodiment will be described. As described above, the number of contents that can be searched for by a user has increased rapidly due to an increase in the capacity of a storage device built in or external to the AV device, or because the AV device is connected to the Internet. Yes.

  However, the user does not grasp all the contents of the search target content. For this reason, it is considered that it is difficult to search for desired content by a method in which a user inputs a keyword and searches for content.

  In the first embodiment, content having relevance with a certain content (hereinafter referred to as reference content) is arranged according to the strength of similarity (relationship). In the first embodiment, the content having a higher degree of association with the reference content is arranged at a position closer to the reference content (see FIGS. 3A and 3B).

  The user selects arbitrary content from the content group arranged according to the degree of association with the reference content. Then, using the selected content as a new reference content, the content is rearranged around the reference content according to the degree of association with the reference content. The user further selects arbitrary content from the newly arranged content group. By repeating this operation, the user searches for desired content.

(Configuration of content search device 1)
FIG. 1 is a diagram illustrating an example of a configuration of a content search apparatus 1 according to the first embodiment. The content search apparatus 1 according to the first embodiment has a configuration in which an antenna 102, a speaker 106, and a monitor 107 are connected to an STB (Set Top Box) 10. The STB 10 includes an HDD 101, a receiving unit 103, a separating unit 104, an AV decoder 105, a CPU 108, a memory 109, a receiving unit 110, a terminal 112, an I / F 113, and a user I / F 114. The receiving unit 110 and the user I / F 114 function as an instruction receiving unit.

  The HDD 101 is a content database that stores a plurality of contents and metadata (content information) of each content. Each content is given a unique number (ID). Instead of the HDD 101, for example, an SSD (Solid State Drive) that is a storage device using a flash memory may be used.

  The HDD 101 stores various contents such as images, videos, sounds, texts, people, books, products, and events. The content information is additional information of each content stored in the HDD 101 and is used for content search. For example, consider a case where a television program by digital broadcasting is recorded (stored) in the HDD 101. In this case, the recorded television program becomes the content. Then, EPG information such as the broadcast time, title, and broadcasting station of the TV program becomes the content information.

  When a photograph taken with a digital camera or the like is stored in the HDD 101, information such as a photograph is content, a title of the photograph, a participant, a photographing date / time, and a photographing place is content information. However, the photo itself may be used in the search in terms of whether the face detection results of the photo match or whether the photo histograms are similar. In this case, the photograph is the content itself, but can also be content information. If the content is a book, information such as the title, author name, and date of issue of the book is the content information.

  Note that among the contents stored in the HDD 101, the moving picture content is pre-assigned to each frame in the form of tag information or the like in units of frames constituting the content. The information (frame information) given to this frame is a group of information indicating characteristics unique to each frame.

  For example, information such as the name of the performer shown in the image included in the frame, information showing the person or scene such as the location, object, state, genre, etc. The chapter title, chapter content, etc. are frame information. Further, recording information in units of frames, program titles, subtitles, genres, content outlines, details, etc. may be used as frame information.

  This frame information may be set by a content creator or distributor (for example, a TV station or a content production company), or a company or third party that provides such a service (for example, a third party). May be set by a user's content viewer, another user of the moving image sharing service, or the like. Moreover, it is good also as an aspect which a user himself inputs.

  The antenna 102 receives a digital broadcast signal transmitted from a broadcast station. The receiving unit 103 selects a desired channel from the digital broadcast signal received by the antenna 102. The receiving unit 103 demodulates the selected digital broadcast signal to generate a TS (transport stream).

  The separation unit 104 separates a broadcast signal (audio ES (audio signal), video ES (video signal)), SI / PSI, and the like from the TS generated by the reception unit 103.

  TS is a multiplexed signal including a broadcast signal and SI / PSI. The broadcast signal is, for example, an MPEG-2 broadcast signal. Broadcast signals are audio ES (Audio Elementary Stream) and video ES (Video Elementary Stream) obtained by encoding video and audio. The PSI is information describing what programs exist in the TS and which programs each ES included in the TS belongs to. The SI includes EPG (Electric Program Guide) information.

  The AV decoder 105 decodes the broadcast signal separated by the separation unit 104 to generate a video signal and an audio signal. In normal times (when playing back without recording), the video signal is output to the monitor 107, and the audio signal is output to the speaker 106. At the time of recording, the video signal and the audio signal are stored in the HDD 101. At this time, EPG information included in the SI / PSIb is stored in the HDD 101 together with the video signal and the audio signal. This EPG information is used as content information when searching for content.

  A CPU (Central Processing Unit) 108 controls the entire broadcast receiving apparatus 1. The memory 109 is composed of ROM, RAM, and the like. The ROM stores a program for the CPU 108 to operate. The RAM is a work area used when the CPU 108 operates.

  FIG. 2 is a diagram illustrating an example of a functional configuration of the content search apparatus according to the first embodiment. When the CPU 108 executes a program stored in the ROM, the search condition acquisition unit 201, the search expression generation unit 202, the search unit 203, the search result analysis unit 204, the relevance calculation unit 205, the arrangement position calculation unit 206, The functions of the display information acquisition unit 207 and the search result presentation unit 208 are realized. The search condition acquisition unit 201, the search expression generation unit 202, the search unit 203, and the search result analysis unit 204 function as an extraction unit. The search expression generation unit 202 functions as a selection unit. Details of each function will be described later.

  The receiving unit 110 receives an operation signal transmitted by radio, such as infrared rays, from the remote controller 20 (hereinafter referred to as the remote controller 20). The remote controller 20 includes various keys necessary for operating the content search apparatus 1 such as a “cursor” key and a “decision” key.

  The user I / F 114 is connected to a key input unit (not shown). The user I / F 114 receives an operation signal from the user operating the key input unit. The key input unit is an operation button of the main body of the content search device 1. Similar to the remote controller 20, the key input unit includes various keys necessary for the operation of the content search device 1.

  The terminal 112 is a terminal to which an external storage device such as a USB memory, an SD memory card, a DVD recorder, or an HDD recorder is connected. An I / F (Inter Face) 114 is an interface for transmitting and receiving data between the storage device connected to the terminal 112 and the CPU 108.

  The user can copy (copy) content from the external storage device connected to the terminal 112 to the HDD 101 by operating the remote controller 20 or the input key. Note that a LAN cable may be connected to the terminal 112 to download content on the Internet to the HDD 101.

  The speaker 106 is an audio output device that outputs an audio signal input from the AV decoder 105. The monitor 107 is a display device that displays a video signal or the like input from the AV decoder 105. Examples of the display device include an LCD and a CRT. The display unit 111 displays a reference content selection screen and a search condition input screen on the monitor 107.

(Standard content selection screen)
FIG. 3A is a diagram illustrating an example of a reference content selection screen displayed on the monitor 107. FIG. 3A shows an example in which a position coordinate system having a conical surface shape is used for the content selection screen. In the first embodiment, a conical surface coordinate system is used.

  In the arrangement slots 301, 302a to 302e, information that best represents the content (hereinafter referred to as representative information) is displayed. For example, if the content is a television program, a title, a broadcasting station, and the like are displayed in the frame together with the image of the television program. If the content is a moving image, information about the moving image is displayed in the frame together with a representative frame. Furthermore, if the content is a book, information such as the title of the book, the author name, and the date of publication is displayed in the frame.

  The arrangement slots 302a to 302e are arranged according to the degree of association with the reference content displayed in the arrangement slot 301. The higher the value on the Z-axis, the higher the relevance content is arranged, and the smaller the value on the Z-axis, the lower the relevance content is arranged.

  In the example shown in FIG. 3A, the content assigned to the placement slots 302a and 302b is highly related to the reference content, and the content assigned to the placement slots 302c to 302e is next related to the reference content. .

  In the first embodiment, content representative information is displayed in a virtual three-dimensional space constituted by a position coordinate system having a conical surface shape shown in FIG. 3A. In this virtual space, the viewpoint (observation viewpoint) to be observed (rendered) can be changed, and the user can set various display forms by changing the viewpoint to observe one coordinate system having a conical surface shape.

  For example, a display form as shown in FIG. 3B can be selected. Further, when the user selects the content display data, the arrangement of the content display data is changed based on the selected content display data. Note that the virtual space in which the content representative information is arranged is not limited to the conical surface coordinate system shown in FIGS. 3A and 3B. For example, it may be displayed in a three-dimensional grid or a two-dimensional grid.

  The user operates the remote controller 20 or the key input unit to select any content from the arranged content groups 302a to 302e. For example, when the content 302a is selected, the selected content 302a is arranged as a new reference content, and other content is rearranged according to the degree of association with the reference content. The user searches for desired content by repeatedly selecting the content displayed on the monitor 107.

  When displaying the screen shown in FIG. 3A, it is necessary to determine the reference content. Various methods can be used to determine the reference content at this time. For example, the viewing state (playback state) of each content stored in the HDD 101 may be monitored, and the currently viewed content may be used as the reference content. When the content is not viewed, the most recently viewed content may be used as the reference content. Further, a list of contents stored in the HDD 101 may be displayed and the user may select.

(Search condition input screen)
FIG. 4A is a diagram illustrating an example of a search condition setting screen displayed on the monitor 107. The search condition is a condition for presenting the searched content to the user. The user operates the remote controller 20 or the key input unit to specify any one of “number of contents”, “degree of association”, or “number of contents and degree of association”. At that time, the user can specify whether or not to perform “stage presentation”. In the example illustrated in FIG. 4A, when “the number of contents and the degree of association” is designated, the user sets both “the number of contents” and “the degree of association”.

  When “number of contents” is specified as a search condition, the user can specify any of the number of contents to be presented, the upper limit of the number of contents, the lower limit of the number of contents, or the allowable range (upper limit and lower limit) of the number of contents. . In the example shown in FIG. 4A, when specifying the allowable range of the number of contents, the user sets both the upper limit and the lower limit of the number of contents.

  Such a search condition specification is suitable for an application in which the number of contents that can be presented is limited or an application that presents a certain number of contents regardless of the degree of relevance. Such applications include, for example, a video database search device.

  When specifying “relevance” as a search condition, the user can specify the relevance of the presented content, the upper limit of the relevance, the lower limit of the relevance, or the allowable range (upper and lower limits) of the relevance. . In the example illustrated in FIG. 4A, when specifying the allowable range of relevance, the user sets both the upper limit and the lower limit of the relevance.

  In general, it is considered that designation (designation of the lower limit) is often performed so that the degree of relevance is a certain numerical value or more. Such setting of search conditions can provide a wide range of contents that can be presented, and is suitable for applications that present contents according to the degree of relevance. Such applications include, for example, a video database search device.

  When “the number of contents and the degree of association” is designated as the search condition, the user sets both the “number of contents” and the “degree of association” for the presented content. The user can specify, for example, “the number of contents is 100 or more and less than 400 and the relevance is 0.6 or more”.

  When “number of contents” is set as a search condition, when “stage presentation” is set, the user designates the number of contents divided into three times such as 10, 30, 60, as shown in FIG. 4A. it can. In this case, 10 content items are presented first, and 30 items in total and 60 items in total are presented in three stages.

  In addition, when “relevance” is set as a search condition, the relevance is divided into three times such that the relevance is 0.8 or higher, the relevance is 0.7 or higher, and the relevance is 0.6 or higher. Can be specified. In this case, content with a relevance level of 0.8 or higher is presented first, content with a relevance level of 0.7 or higher, and finally content with a relevance level of 0.6 or higher are presented in three stages. The number of presentations is not limited to three, and an arbitrary number can be set.

  In other words, when “stage presentation” is selected, search results that satisfy the search conditions set by the user can be presented in several stages, instead of presenting the search results at once. Conditions can be given. When “stage search” is selected as the search condition, the user instructs to perform the remaining search after the search result is presented. Further, if the set search condition is satisfied, the remaining search can be canceled. When the search condition is satisfied, for example, when “number of contents” is set as the search condition, the number of cases equal to or less than the set number of contents is searched. Further, for example, when “relevance” is set as a search condition, a search result equal to or higher than the set relevance is obtained.

  In addition, the user searches for content by specifying a search method. In the first embodiment, first to fifth search methods are prepared as content search methods. Each search method is given a unique number (ID). The user designates the search method by designating this ID (No. 1 to No. 5). The search method may be specified as an option. In this case, the default search method is automatically specified.

  FIG. 4B is a diagram illustrating an example of a search condition setting screen displayed on the monitor 107. FIG. 4B is an example of a screen for setting the condition element and the weighting factor of the search expression generated by the search expression generating unit 202. Conditional elements and weighting factors will be described later.

In the example shown in FIG. 4B, “title”, “performer”, “program content”, “genre”, “broadcast time”, and “broadcast station” are set as condition elements. Further, the weighting factors a ₀ to a ₅ of each condition element are set to 0.5, 0.3, 0.1, 0.05, 0.03, and 0.01. The setting of the condition element and the weighting factor may be optional.

  In the content search apparatus 1 according to the first embodiment, the search condition can be set by the user using the search condition setting screen as described above. Therefore, desired contents can be obtained more efficiently by flexibly operating the search conditions while viewing the search results presented on the screen described in FIG. 3A.

(Get search conditions)
As described with reference to FIG. 3A, when the user selects arbitrary content from the content group displayed on the monitor 107, the search condition acquisition unit 201 acquires content information corresponding to the content selected by the user from the HDD 101. .

(Generation of search expression)
The search expression generation unit 202 generates a content search expression based on the content information acquired by the search condition acquisition unit 201. Hereinafter, generation of a search expression will be described, but before that, a method of calculating the degree of association will be described.

(Calculation of relevance)
In the first embodiment, the relevance level RC with respect to the reference content is calculated, and the content is arranged according to the relevance level. The degree of association RC between contents is calculated using the following equation (1).
_{RC = a 0 x 0 + a} 1 x 1 + ... + a n x n ... (1)
In the formula (1), (x ₀ , x ₁ ,..., X _n ) is a conditional element (item). Further, (a ₀ , a ₁ ,..., _An ) are weighting factors.

(Condition element x)
The condition elements (x ₀ , x ₁ ,..., X _n ) are minimum units of search conditions. For example, when the user selects a TV program as content in FIG. 3, the “title”, “performer”, “program content”, “genre”, “broadcast time”, “broadcast station”, etc. of the selected content are displayed. Set as a condition element.

  When a book is selected as the content, “title”, “author name”, “keyword”, “publisher”, “classification”, “ISBN (International Standard Book Number)”, “ISSN (International Standard Book Number)” Serial Number) ”and“ Publishing Date ”are set as condition elements.

Each condition element x ₀ to x _n is represented by a Boolean value of 1 or 0. That is, if there is a condition element that matches the content information of the content selected by the user, the value of the condition element is 1, and if there is no condition element that matches, the value is 0.

(Weight coefficient a)
The weighting factors (a ₀ , a ₁ ,..., A _n ) are parameters for weighting each condition element. A higher coefficient is set for a condition element that is considered to be more important to the user. In the first embodiment, to a _n weighting coefficient a _0, the following equation (2) is set to satisfy the equation (3).
a ₀ + a ₁ +... + a _n = 1 (2)
a ₀ > a ₁ >...> a _n (3)

For example, consider a case where a desired television program is searched for as content stored in the HDD 101. In this case, a condition element is set based on EPG information as content information. In the case of a television program, in general, it is considered that “title”, “performer”, “program content”, “genre”, “broadcast time”, and “broadcast station” are in descending order of importance. In this case, the weighting factors a ₀ to a ₅ of each component can be set to 0.5, 0.3, 0.1, 0.05, 0.03, and 0.01, for example.

Note that the value of the weight coefficient is an example. In the first embodiment, is set sum to one to each weighting coefficient a ₀ not a _n, it may be set range different for each weighting coefficient. In addition, the search formula does not necessarily have to have a form in which weights are added. What is necessary is just to be able to calculate the relevance for determining whether or not the search result is appropriate.

  As the weighting factor for each constituent requirement, for example, the priority of the constituent element and the weighting factor for the priority may be determined in advance for each genre of content and stored as table data. Then, the content genre is determined from the content information of the content selected by the user, and the weighting coefficient of each condition element is set by referring to the table data.

  As described above, in the first embodiment, content information corresponding to the reference content is set as a condition element. A weighting factor for each condition element is set according to the priority. Then, for each content stored in the HDD 101, the relevance RC for the selected content is calculated by calculating equation (1) in which the condition element and the weighting coefficient are set.

However, there is a wide variety of information acquired as content information here. Furthermore, the number of contents stored in the HDD 101 is enormous. Under such conditions, the match / mismatch of each condition element (x ₀ , x ₁ ,..., X _n ) is determined for all the contents stored in the HDD 101, and the relevance of equation (1) If RC is calculated, the amount of calculation becomes enormous and it is not realistic.

Here, a value is added to the expression (1) if the condition elements match. For this reason, it is possible to know the minimum value of the relevance of the expression (1) that is finally calculated without determining whether the condition elements match or not. For example, if the boolean value of x ₀ is 1, the degree of association equal to or greater than the value of the weighting factor a _0. Moreover, the relevance RC if boolean _{x 1} is 1, becomes equal to or greater than the value of the weighting coefficients _{a 1,} x 0, if both 1 Boolean _{x 1,} relevance RC weighting factor _a 1, It becomes a a ₀ total value or more.

  That is, the number of contents related to the reference content can be narrowed down to the search condition, that is, the number of contents to be presented to the user, without calculating the relevance level RC of equation (1) for all the contents.

  For example, consider a case where “title”, “performer”, “program content”, “genre”, “broadcasting time”, and “broadcasting station” are set as condition elements when calculating the relevance RC. In this case, the number of contents can be narrowed down by determining match / mismatch for any of the five items without determining match / mismatch for all the five items.

  Then, after narrowing down the number of contents until the search condition is satisfied or a value close to the search condition, the relevance RC is calculated by determining whether all the condition elements match or not with respect to the narrowed content. If the relevance level RC is calculated after narrowing down the content in this way, the amount of calculation during content search can be reduced, and as a result, efficient search can be performed.

(Content search method)
FIG. 5 is a diagram illustrating an example of the relationship between the number of contents and the degree of association RC. As can be seen from the configuration of equation (1), the number of matching condition elements is large when the relevance level RC is large, and the number of matching condition elements is small when the relevance level RC is small. That is, when the relevance level RC is large, the search condition becomes severe, and when the relevance level RC is small, the search condition becomes loose.

  For this reason, generally, if the relevance level RC when searching for content is large, the number of content items to be searched tends to decrease, and if the relevance level RC is small, the number of content items to be searched tends to increase. Based on the above description, the content search method in the first embodiment will be described below.

(First search method)
The first search method is a method in which a search is performed with a search formula having a high degree of relevance RC, and content is searched while decreasing the value of the relevance RC until the number of searched contents reaches the number of contents presented to the user. is there.
This first search method is effective when the relevance level RC of the content satisfying the search condition is relatively high. In other words, it is effective when the number of contents presented to the user is relatively small.

For example, consider using the following formula (4) as a search formula.
_{RC = a 0 x 0 + a} 1 x 1 + a 2 x 2 ... (4)
(X ₀ , x ₁ , x ₂ ) are condition elements. (A ₀ , a ₁ , a ₂ ) are weighting factors.

Here, in order for the relevance RC in the equation (4) to be a value other than 0, the value of (x ₀ , x ₁ , x ₂ ) need not be (0, 0, 0). Therefore, as the search expression in which the relevance RC of the expression (4) is a value other than 0, the following seven patterns of the following expressions (5) to (11) are conceivable.

Pattern 1: RC = a ₀ x ₀ + a ₁ x ₁ + a ₂ x ₂ (5)
Pattern _{2: RC = a 0 x 0} + a 1 x 1 ... (6)
Pattern 3: RC = a ₀ x ₀ + a ₂ x ₂ (7)
Pattern 4: RC = a ₀ x ₀ (8)
Pattern 5: RC = a ₁ x ₁ + a ₂ x ₂ (9)
Pattern 6: RC = a ₁ x ₁ (10)
Pattern 7: RC = a ₂ x ₂ (11)

Considering that the weighting factors (a ₀ , a ₁ , a ₂ ) have a relationship of a ₀ > a ₁ > a ₂ , the relevance level RC is highest in the case of pattern 1. Therefore, in the first search method, content is searched using the search formula in the case of pattern 1. If the search result does not match the search condition, the content is searched using the search formula in the case of pattern 2, and the relevance RC gradually with pattern 3 and pattern 4 until the number of contents finally matches the search condition. Narrow the content by repeating the search using a low-level search expression.

(Second search method)
The second search method is a method in which a search is performed with a search formula having a low relevance level RC, and content is searched for while increasing the relevance level RC until the number of content items that match the search conditions is obtained. This second search method is effective when the relevance RC of the content satisfying the search condition is relatively low. In other words, it is effective when the number of contents presented to the user is relatively large.

  In the second search method, the content is searched using the search formula in the case of the pattern 7 having the lowest relevance level RC. If the search result does not match the search condition, the content is searched using the search formula in the case of pattern 6 and gradually associated with patterns 5 and 4 until the number of contents finally matches the search condition. The content is narrowed down by repeating the search using a search formula with a high degree RC.

(Third search method)
The third search method is a method of searching for content using a search formula of the relevance level RC that is closest to the median value of possible relevance levels RC, and thereafter searching for content in the manner of binary search. This search method can suppress variations in time (standard deviation) until a final search result is obtained.

  In the first embodiment, since the maximum value of the relevance level RC is “1”, the content is searched using a search formula that makes the relevance level RC closest to “0.5”. When searching for content using this search formula, if the number of searched content does not satisfy the search condition, it is necessary to increase the number of searched content. In order to increase the number of contents to be searched, it is necessary to lower the relevance RC. Therefore, the content is narrowed down by searching for the content using a search formula having a relevance level RC that is closest to the intermediate value “0.25” of “0” and “0.5”.

  Further, when the number of searched contents exceeds the search condition, it is necessary to reduce the number of searched contents. And in order to reduce the number of contents searched, it is necessary to raise the relevance RC. Therefore, the content is narrowed down by searching for the content using a search formula that has a relevance level “0.5” and a value closest to the intermediate value “0.75” of “1”.

(Fourth search method)
The fourth search method is a method for searching for content by function fitting. This fourth search method is effective when the number of condition elements is large and the relevance level RC can take a wide range of values, or when the relationship between the relevance level RC and the number of searched contents is complex.

  In the fourth search method, first, content is searched using three search formulas in which the value of relevance RC is the highest value, the lowest value, and the intermediate value. Then, a relationship (function = P (RC)) between the value of the relevance RC and the number of contents is calculated by fitting.

  Then, the relevance level RC corresponding to the number of contents designated by the search condition is calculated from the function = P (RC) obtained by the fitting. Next, the content is searched using a search expression that matches or is closest to the calculated relevance RC. If the number of contents searched when using the search formula does not match the search condition, the relation between the value of relevance RC and the number of contents is further added by adding the result of the previous search (function = P (RC)) Is calculated by fitting. Then, the relevance level RC is calculated using the newly calculated function, and a search expression that is the closest value to the relevance level RC is calculated to search for the content. By repeating such an operation until the number of contents matching the search condition is searched, the contents are narrowed down.

(Fifth search method)
Search conditions can also be set with the relevance RC. In this case, a search expression having the set relevance level RC is generated to search for content. For example, when the relevance level RC is set as 0.6 as a search condition, a search formula is generated for a combination of condition elements that has a relevance level RC of 0.6 or more.

  The search unit 203 searches the content stored in the HDD 101 using the search formula generated by the search formula generation unit 202. The search unit 203 collates the search formula condition elements (items) generated by the search formula generation unit 202 with the content information items of the content stored in the HDD 101 one by one. The search unit 203 only determines whether or not the condition elements between the contents match, and does not calculate the final relevance. The relevance level of the content is calculated by a relevance level calculation unit 205 described later.

  For example, if the content is text, full-text matching, partial matching, matching with dictionary data (including ontology database, thesaurus database, etc.), etc. are examined to determine whether each condition element matches or does not match. Also, if the content is an image, video, audio, etc., matching information obtained from the image, video, audio, such as color information contained in the image, video, audio, number of pixels, aspect ratio, number of frames, histogram, waveform data, etc. By matching, it is determined whether each condition element matches or does not match.

  In addition, when a condition based not only on the relevance level RC but also on the number of contents is specified as a search condition, the search expression generation method when the search condition shown above is specified on the number of contents is also used. Thus, a search expression is generated.

(Analysis of search results)
The search result analysis unit 204 determines whether the search result of the search unit 203 satisfies the search condition. When the search result of the search unit 203 does not satisfy the search condition, the search result analysis unit 204 instructs the search formula generation unit 202 to generate the next search formula. When the search result of the search unit 203 satisfies the search condition, the search result analysis unit 204 instructs the relevance calculation unit 205 to calculate the relevance RC with respect to the reference content for each searched content.

(Exception handling)
Depending on the set search condition, there is a possibility that a contradiction occurs between the search condition and the search result. Therefore, in the first embodiment, when a contradiction occurs between the search condition set by the user and the search result, the search condition setting screen is displayed on the monitor 107, and the user is made to input the search condition again. The user can obtain desired content more efficiently by flexibly operating the search conditions while viewing the presented search results.

  Hereinafter, a case where a contradiction occurs in the presentation result will be described. When the following case occurs, the search result analysis unit 204 instructs the display unit 111 to display the search condition input screen described with reference to FIGS. 4A and 4B on the monitor 107. The user resets the search conditions on the search condition input screen displayed on the monitor 107.

(First exception)
When “the number of contents and the degree of relevance RC” is designated as the search condition, it is possible to narrow down the contents more strictly, but there is a possibility that a contradiction occurs between the search result and the search condition. For example, when the number of contents is set to “100 or more and 400 or less” and the “relevance RC” is set to “0.6 or more” as the search condition, the number of contents with a relevance RC of 0.6 or more is 100. It is thought that it is less than. Moreover, the case where the number of contents whose relevance RC is 0.6 or more exceeds 400 is considered.

(Second exception)
As shown in FIG. 5, there is a general tendency that the number of contents is small when the relevance level RC is large, and the content number is large when the relevance level RC is small. However, there is a case where the relationship between the degree of relevance RC and the number of contents is locally reversed. FIG. 6 is a diagram showing an example of the relationship between the number of contents and the relevance level RC in such a case.

  As shown in FIG. 6, when the tendency changes locally, a search formula with a lower relevance RC is used if the number of contents in the search result is small, and a search formula with a higher relevance RC if the number of contents in the search result is large. If you repeat the process of using, you may fall into an infinite loop.

(Third exception)
In addition, when “stage presentation” is designated as the search condition, care must be taken in specifying the search method. For example, the relevance RC is specified such that the relevance RC is 0.8 or higher first, the relevance RC is 0.7 or higher, and finally the relevance RC is 0.6 or higher. Consider a case in which a condition for presenting content with a relevance RC of 0.6 or more is given.

  In this case, the search expression may be generated in the same manner as when the search condition is specified by the relevance level RC. When the number of contents is specified such as first 10 cases, then 30 cases, and finally 60 cases, and conditions are given so as to finally present 100 contents in three stages, the search condition is It is necessary to generate the search formula by the first method when it is specified by the number of contents.

For example, as shown in FIG. 6, when the number of contents is locally small (part of a ₁ + a ₂ ), the contents are searched by the second search method (a search expression is generated from the one having the lower relevance RC). Think about it. In this case, a reverse phenomenon occurs in which the number of contents presented when the relevance level RC is low is smaller than the number of contents presented when the relevance level RC is high.

  Referring to FIG. 6, after presenting the search result by the above-described search formula (9), the search result by the search formula (8) or (7) is presented. Then, the degree of relevance RC is higher for the content presented later than the content presented earlier. Even when the search condition is specified by the number of contents, it is necessary to search for contents having a high degree of relevance RC. In some cases, the same caution is required even when step-by-step presentation is not specified.

  When such an exception occurs, the search condition setting screen described with reference to FIGS. 4A and 4B is displayed on the monitor 107, and the user is allowed to reset the search condition. For example, when the first exception occurs, the re-search can be performed so that the number of contents is 100 or less or less than 400 by resetting the relevance level RC to a smaller value.

  If the second exception occurs, the number of times the search expression is generated is counted, and the process is interrupted when the count reaches a predetermined number, or the search result at that time is retained. Above, you may make it escape from the infinite loop processing by changing the retrieval formula.

  As described above, when there is a contradiction between the search condition and the search result, it is possible to efficiently search for desired content from the HDD 101 in which a large number of contents are stored by allowing the user to re-enter the search condition.

(Calculation of relevance RC)
Based on an instruction from the search result analysis unit 204, the relevance level calculation unit 205 calculates the relevance level RC with the reference content for each content searched by the search unit 203. The relevance calculation unit 205 determines the reference content C1 that is a calculation target of the relevance RC. Next, content information C11 associated with the reference content C1 is acquired from the HDD 101.

  Subsequently, when the degree-of-association calculation unit 205 determines the content C2 that is the target of calculating the degree of association RC, the degree-of-association calculation unit 205 acquires content information C21 associated with the content C2 from the HDD 101. Next, the relevance calculation unit 205 calculates the relevance RC of the content C2 with respect to the reference content C1 using the above-described equation (1).

The information acquired as content information here is diverse, but for the sake of simplicity, the content information is composed of five sets of {recording date and time, content title, channel, genre, performer list}. Assume that these five pieces of content information are conditional elements x ₀ to x ₄ from the top.

Then, the relevance level RC of C2 with respect to the content C1 can be expressed by a weighted linear sum of the following 5 relevance levels RCRC1 to RC5.
RC1 ₌ a _{0 x} 0 (recording date and time, recording date and time of content information C21 of content information C11)
RC2 = a ₁ x ₁ (title of content information C11, title of content information C21)
RC3 = a ₂ x ₂ (channel of content information C11, channel of content information C21)
RC4 = a ₃ x ₃ (genre of content information C11, genre of content information C21)
RC5 = a ₄ x ₄ (Performer list of content information C11, Performer list of content information C21)

  In the above description, the method for calculating the degree of association RC between contents using the contents information corresponding to the contents as the contents information has been described. However, in the case of moving image content, face recognition information, a luminance histogram, or the like may be used to calculate the relevance level RC. Further, the frame information described above may be used.

(Arrangement position of arrangement slot)
The arrangement position calculation unit 206 arranges the arrangement slot in a virtual three-dimensional space (hereinafter referred to as a virtual space). The arrangement position of the arrangement slot is determined according to the number of contents presented to the user and the relevance level RC calculated by the relevance level calculation unit 205.

  As shown in FIG. 3A, the placement slot of the reference content is placed at the origin (x, y) = (0, 0) of the virtual space. Then, an arrangement slot for another content is arranged with the normal direction of the content as the z-axis (normal vector = (0, 0, 1)).

  At this time, the arrangement position calculation unit 206 determines the arrangement position on the Z axis in the virtual space based on the association degree RC calculated by the association degree calculation unit 205. Specifically, as shown in FIG. 3A, a conical arrangement coordinate system is provided in the virtual space, and the position of the truncated cone on which the arrangement target content is arranged is determined on the Z axis. Here, the center of the truncated cone is orthogonal to the Z-axis, and as the value on the Z-axis (Z value) is larger, the relevance level RC is larger, and as the Z value is smaller, the relevance level RC is smaller.

  Note that content information, for example, genre information, may be used when determining the placement position of the placement slot. For example, the center of FIG. 3B is the origin (x, y) = (0, 0), and the genre is set in each azimuth direction on the XY plane centered on the origin (x, y) = (0, 0). Types are assigned in advance. Based on the genre information included in the content information of the content to be arranged, the arrangement position on the XY plane in the virtual space is determined. Specifically, the genre information of the content to be arranged is compared with the genre information assigned to each direction on the XY plane, and the coincident direction is determined as the arrangement position on the XY plane.

The allocation to each direction is not necessarily fixed, and may be dynamically changed according to the environment at the time of use. For example, referring to a user's past operation history (not shown), it may be assigned to each direction in the order of the genre of a program that is frequently viewed. Moreover, it is good also as changing the allocation to each azimuth | direction by the instruction information input via the operating device from a user. Furthermore, the information allocated to each direction is not necessarily one piece of information. For example, a mode in which a plurality of types of information are simultaneously allocated, such as a genre in one direction and a recording date and time in another direction. It is good.
(Get display information)

  The display information acquisition unit 207 acquires the content representative information searched by the search unit 203 from the content information stored in the HDD 101.

(Presentation of search results)
The search result presentation unit 208 generates an image in which the representative information of the content corresponding to the arrangement slot is displayed in the arrangement slot arranged in the virtual space by the arrangement position calculation unit 207. Then, this image is displayed on the monitor 107 and the search result is presented to the user.

(Operation of Content Search Device 1)
FIG. 7 is a flowchart showing an example of the operation of the content search apparatus 1 according to the first embodiment. Hereinafter, the operation of the content search apparatus 1 according to the first embodiment will be described.

  The search condition acquisition unit 201 acquires content information and search conditions corresponding to the content selected by the user (step S101).

  The search formula generation unit 202 generates a search formula based on a search method specified by the user or set as a default. (Step S102).

  The search unit 203 searches the content group stored in the HDD 101 for content that satisfies the search formula generated by the search formula generation unit 202 (step S103).

  The search result analysis unit 204 determines whether the search result in the search unit 203 satisfies the search condition acquired by the search condition acquisition unit 201 (steps S104 and S105).

  When the search result satisfies the search condition, the search result analysis unit 204 instructs the relevance calculation unit 205 to calculate the relevance level RC with respect to the reference content for each searched content (Yes in step S105). .

  Based on the instruction from the search result analysis unit 204, the relevance calculation unit 205 calculates the relevance RC with the reference content for each content searched by the search unit 203 (step S106).

  The arrangement position calculation unit 206 arranges the arrangement slot in the virtual space according to the relevance level RC calculated by the relevance level calculation unit 205 (step S107).

  The display information acquisition unit 207 acquires the content representative information searched by the search unit 203 from the content information stored in the HDD 101 (step S108).

  The search result presentation unit 208 generates an image in which the representative information of the content corresponding to the arrangement slot is displayed in the arrangement slot arranged in the virtual space by the arrangement position calculation unit 207. Then, this image is displayed on the monitor 107 and the search result is presented to the user (step S109).

  If the search result does not satisfy the search condition (No in step S105), the search result analysis unit 204 determines whether there is a contradiction in the presentation result, that is, whether exception processing is necessary (step S110).

  When exception processing is not necessary (No in step S110), the search result analysis unit 204 instructs the search formula generation unit 202 to regenerate the search formula. The search expression generation unit 202 regenerates the search expression in response to an instruction from the search result analysis unit 204. At this time, the search expression generation unit 202 regenerates the search expression based on a search method designated by the user or set as a default.

  If there is a contradiction in the presentation result (Yes in step S110), the search result analysis unit 204 instructs the display unit 111 to display the search condition input screen described with reference to FIGS. 4A and 4B. The user inputs search conditions, and the search condition acquisition unit 201 acquires the search conditions input by the user (step S101).

  The search expression generation unit 202 or the search result analysis unit 204 repeats the operations from step S101 to S105 and step S110 until the search result satisfies the search condition.

  As described above, the content search device 1 according to the first embodiment calculates the relevance RC for the extracted content after extracting the content that satisfies the search condition. For this reason, the amount of calculation at the time of content search can be reduced, and high-speed search can be performed. Further, the content satisfying the search condition is displayed according to the degree of association with the reference content. Since the user can search for the content while checking the displayed content, the user can more easily search for the desired content.

  Further, the requested content can be obtained more efficiently by flexibly manipulating the search conditions while looking at the presented search results. Furthermore, since the user can set search conditions, the user can search for desired content more efficiently by operating the search conditions while checking the search results displayed on the monitor 107. Note that the content satisfying the search condition may be simply displayed as a list on the monitor 107. In this case, although the content is not arranged according to the degree of relevance, the content satisfying a certain condition is displayed on the monitor 107, so that the user can effectively search for the desired content.

(Second Embodiment)
FIG. 8 is a diagram illustrating an example of a configuration of a content search apparatus according to the second embodiment. FIG. 9 is a diagram illustrating an example of a functional configuration of the content search apparatus according to the second embodiment. 10A and 10B are diagrams illustrating an example of information (search history) stored in the search history (extraction history) DB. In the second embodiment, an embodiment in which a search history (extraction history) is stored and this search history is used will be described.

(Configuration of content search device 2)
First, the configuration of the content search device 2 according to the second embodiment will be described. As shown in FIG. 8, the content search apparatus 2 according to the second embodiment has a configuration in which an antenna 102, a speaker 106, and a monitor 107 are connected to an STB (Set Top Box) 10. The STB 10 includes an HDD 101, a receiving unit 103, a separating unit 104, an AV decoder 105, a CPU 108, a memory 109, a receiving unit 110, a terminal 112, an I / F 113, a user I / F 114, and an HDD 115. The receiving unit 110 and the user I / F 114 function as an instruction receiving unit.

  FIG. 9 is a diagram illustrating an example of a functional configuration of the content search apparatus according to the second embodiment. Similar to the first embodiment, the CPU 108 executes a program stored in the ROM, whereby the search condition acquisition unit 201, the search expression generation unit 202A, the search unit 203, and the search result analysis unit 204 shown in FIG. The functions of the association degree calculation unit 205, the arrangement position calculation unit 206, the display information acquisition unit 207, the search result presentation unit 208, the search history update unit 209, and the search history acquisition unit 210 are realized. The search condition acquisition unit 201, the search expression generation unit 202A, the search unit 203, and the search result analysis unit 204 function as an extraction unit. The search expression generation unit 202A functions as a selection unit. In addition, about the component same as the component demonstrated in FIG. 1, FIG. 2, the same code | symbol is attached | subjected and the overlapping description is abbreviate | omitted.

  The HDD 115 is a search history database that stores search results in the search unit 203 as a history. As shown in FIG. 10A, the HDD 115 stores the reference content ID, the search formula used for searching the content, the number of content searched by the search formula, and the ID of the searched content in association with each other. The HDD 115 may be configured to store the analysis result in the search result analysis unit 204 as a search history.

  By storing the analysis result in the search unit 203 in the HDD 115, it is not necessary to narrow down the content when the same content is selected as the reference content. For this reason, it is possible to effectively reduce the amount of calculation when narrowing down the content, and it is possible to search the content efficiently.

  Even if the search formulas are not exactly the same, if you want to search with a subset of the search formulas that are already stored, you can search for extra content by searching the search history as a mother set. More efficient processing can be realized without including.

  As shown in FIG. 10B, the search result in the search unit 203 may be stored in the form of an inverted index. When storing the search history in the form of an inverted index, the number of contents and the content ID are stored for each condition element. By storing the search history in the inverted index format, the number of contents searched for each condition element can be estimated.

For example, the search history shown FIG. 10B, it is understood that the number of contents condition elements X ₁ match is stars 512, also it can be seen that there review content number 1194 condition element X ₂ are identical. For example, if the search condition is 1000, if using the X ₂ as a condition element of the search expression, it is possible to obtain the search results that meet faster search. As described above, in the second embodiment, when the search expression generation unit 202 generates a search expression, the search history stored in the HDD 115 is used.

(Update search history)
The search history update unit 209 stores the search result in the search unit 203 in the HDD 115 as a search history. The search history update unit 209 stores the analysis result obtained by the search unit 203 after converting it into a data format that can be stored in the HDD 115.

  When storing the search result in the search unit 203 in the HDD 115, the search history update unit 209 determines whether the search result in the search unit 203 is already stored in the HDD 115. When the search result is stored in the HDD 115, the search history update unit 209 does not update the search history in the HDD 115. When the search result is not stored in the HDD 115, the search history update unit 209 updates the search history in the HDD 115.

For example, in the case where the result of the search using a search expression such as x ₀ & X ₁ is stored as a transposed index, in this search, content that matches both the x ₀ condition element and the x ₁ condition element can be searched. . Therefore, the search history update unit 209 performs additional registration in the content ID lists of both condition elements and updates the number of contents.

(Get search history)
The search history acquisition unit 210 acquires a search history from the HDD 115 based on an instruction from the search expression generation unit 202A.

(Generation of search expression)
The search expression generation unit 202A generates a search expression using the search history of the HDD 115 in order to generate a search expression that can obtain a search result close to the search condition. The method of using the search history differs depending on which search method is selected from the first to fourth search methods. Hereinafter, generation of a search expression using a search history will be described.

(First search method)
The first search method is a method for searching for content while decreasing the value of the relevance RC until the number of contents to be presented to the user is reached. When the first search method is specified, the search expression generation unit 202A generates a search expression that is closest to the search condition and has a small number of contents to be searched, from the search history acquired by the search history update unit 209. To do.

For example, if the number of contents specified by the search condition is 600, the search expression generation unit 202A generates the following expression (15) from the search history shown in FIG. 10A. Since the subsequent search method is the same as the first search method described in the first embodiment, a duplicate description is omitted.
E = a ₁ x ₁ (15)

(Second search method)
The second search method is a method for searching for content while increasing the value of the relevance RC. When the second search method is specified, the search expression generation unit 202A generates a search expression that is closest to the search condition and has a large number of contents to be searched, from the search history acquired by the search history update unit 209. To do.

For example, if the number of contents specified by the search condition is 600, the search expression generation unit 202A generates the following expression (16) from the search history shown in FIG. 10A. Since the subsequent search method is the same as the second search method described in the first embodiment, a duplicate description is omitted.
E = a ₂ x ₂ (16)

(Third search method)
The third search method is a method for searching for content in the manner of binary search. When the third search method is designated, the search expression generation unit 202A searches for content in the manner of a binary search from the search history acquired by the search history update unit 209.

  For example, if the number of contents specified by the search condition is 600, the search expression generation unit 202A, from the search history shown in FIG. 10A, the relation of the expression (15) that is closest to the number of contents to be searched is 600. A search expression that is an intermediate value between the degree RC and the degree of association RC of the above expression (16) is generated. Since the subsequent search method is the same as the third search method described in the first embodiment, a duplicate description is omitted.

(Fourth search method)
The fourth search method is a method for searching for content by function fitting. When the third search method is designated, the search expression generation unit 202A calculates a function = P (E) to be fitted to the search history acquired by the search history update unit 209. Since the subsequent search method is the same as the fourth search method described in the first embodiment, a duplicate description is omitted.

In addition, when the search unit 203 performs a search using the search formula generated by the search formula generation unit 202, the search formula may be a subset of the search formula that has already been searched. In such a case, the search formula generation unit 2 issues a command to the search history update unit 209 to acquire from the HDD 115 the search result of the search formula that has already been searched. The search history update unit 209 acquires the search history from the HDD 115 in accordance with the above command.
In the search expression generation unit 202A, the generated search expression and the search result of the search expression that has already been acquired and are acquired are transmitted to the search unit 203. Then, the search unit 203 obtains a search result for the newly generated search expression using the search result of the search expression that has already been searched as a mother set.

(Operation of Content Search Device 2)
FIG. 11 is a flowchart showing an example of the operation of the content search apparatus 2 according to the second embodiment. The operation of the content search device 2 according to the second embodiment will be described below.

  The search condition acquisition unit 201 acquires content information and search conditions corresponding to the content selected by the user (step S201).

  The search history update unit 209 acquires a search history from the HDD 115 based on an instruction from the search formula generation unit 202A (step S202). Here, the search history is acquired from the HDD 115 according to the search condition acquired in step S201.

  The search expression generation unit 202A generates a search expression based on the search condition acquired by the search condition acquisition unit 201 and the search history acquired by the search history update unit 209. (Step S203). In the second embodiment, a search expression is generated according to the search history acquired in step S201.

  The search unit 203 searches the content group stored in the HDD 101 for content that satisfies the search formula generated by the search formula generation unit 202 (step S204).

  The search history acquisition unit 210 stores (updates) the search result in the search unit 203 in the HDD 115 (step S205). When the search result is stored in the HDD 115, the search history update unit 209 does not update the search history in the HDD 115.

  Next, the content search device 2 performs the processing (first processing) from step S104 to step S110 described in FIG. 7 (step S206). The operation of the content search apparatus 2 described with reference to FIG. 9 is performed until the search result satisfies the search condition.

  As described above, the content search device 2 according to the second embodiment stores the search result of the search unit 203 as a search history in the HDD 115, and generates a search formula using this search history. Therefore, it is possible to generate a search expression that can obtain a search result close to the condition specified by the search condition. As a result, content can be narrowed down more efficiently than in the first embodiment. Other effects are the same as those of the first embodiment.

(Third embodiment)
In the third embodiment, an embodiment will be described in which a search history of updated content is created in advance when the content of the HDD 101 is updated. A search can be performed more efficiently by creating a search history of updated content in advance.

  FIG. 12 is a diagram illustrating an example of a functional configuration of the content search apparatus according to the third embodiment. The configuration of the content search device 3 according to the third embodiment is the same as the configuration of the content search device 2 according to the second embodiment described with reference to FIG.

  Similar to the first embodiment, the CPU 108 executes a program stored in the ROM, whereby the search condition acquisition unit 201, the search expression generation unit 202B, the search unit 203, and the search result analysis unit 204 shown in FIG. The functions of the association degree calculation unit 205, the arrangement position calculation unit 206, the display information acquisition unit 207, the search result presentation unit 208, the search history update unit 209, the search history acquisition unit 210, and the content update unit 211 are realized. Here, the search condition acquisition unit 201, the search expression generation unit 202B, the search unit 203, and the search result analysis unit 204 function as an extraction unit. The search expression generation unit 202B functions as a selection unit. In addition, about the component same as the component demonstrated in FIG. 1, FIG. 9, the same code | symbol is attached | subjected and the overlapping description is abbreviate | omitted.

(Content update)
The content update unit 211 updates the content of the HDD 115 that is a content database. The content update unit 211 stores content update information together when content is updated. For example, the update flag and the update date / time are stored in association with the content.

  If the content is stored in the HDD 101 during the recording of the TV program, the content update unit 211 stores an update flag and an update date and time together with the content. In addition, when the user operates the remote controller 20 or the input key to copy (copy) content on the Internet from the storage device or LAN cable connected to the terminal 112 to the HDD 101, the content update unit 211 updates an update flag or update. The date and time are stored together with the content.

  When the content is a television program, the television program is automatically recorded or downloaded (updated) from the terrestrial broadcast, satellite broadcast or network received by the antenna 102. In addition, when the content is data of a manufactured or sold product, the content update unit 211 updates the content of the HDD 115 in synchronization with the product group manufacturing system or sales system.

(Create search history)
When the content of the HDD 101 is updated, the search expression generation unit 202B, the search unit 203, and the search history update unit 209 create a search history using this content as reference content. The search history of the updated content is created in the background.

(Create search history)
FIG. 13 is a diagram illustrating an example of the operation of the content search apparatus 3 according to the third embodiment. Hereinafter, creation of a search history by the content search device 3 will be described. When the content stored in the HDD 101 is updated, the search expression generation unit 202B generates a search expression (step S301).

For example, when the content ID of the updated content is 100 and the condition elements of this content are x ₀ , x ₁ , x ₂ , the search expression generation unit 202B uses the following expressions (17) to (21): Generate search expressions sequentially.

E = a ₀ x ₀ + a ₁ x ₁ + a ₂ x ₂ (17)
_{E = a 0 x 0 + a} 1 x 1 ... (18)
E = a ₀ x ₀ + a ₂ x ₂ (19)
E = a ₀ x ₀ (20)
E = a ₁ x ₁ + a ₂ x ₂ (21)
E = a ₁ x ₁ (22)
E = a ₂ x ₂ (23)

The search formulas may be generated in any order. For example, it may be generated from a search expression having a high relevance level RC as in the first search method, or may be generated from a search expression having a low relevance level RC as in the second search method. Alternatively, a search expression (expression (20), expression (22), expression (23)) may be generated only for the condition elements x ₀ , x ₁ , x ₂ .

  The search unit 203 searches the HDD 101 for content that satisfies the search formula generated by the search formula generation unit 202B (step S302). The search history update unit 209 stores the search result in the search unit 203 in the HDD 115. As described with reference to FIG. 10A, the HDD 115 stores the reference content ID, the search formula used for searching for the content, the number of content searched by the search formula, and the ID of the searched content in association with each other. Further, as described in FIG. 10B, the search history in the search unit 203 may be stored in the form of an inverted index.

  The search expression generation unit 202B generates a search expression (expression (17) to expression (23)) for all combinations of condition elements for the updated content. Then, the process from step S301 to step S303 is performed until the search result in the search unit 203 is stored in the HDD 115 (No in step S304).

  The content updated at regular intervals may be searched to generate a search history of the searched content. In this case, there is a possibility that a search expression is already generated for the updated content and the search result is stored in the HDD 115. In such a case, the search expression generation unit 202B generates a search expression other than the search results stored in the HDD 115. The search unit 203 searches the content in the HDD 101 according to the search formula generated by the search formula generation unit 202B. The search history update unit 209 stores the search result in the search unit 203 in the HDD 115 and updates the search history.

  For example, if the search history related to the condition elements x1 and x2 is not stored in the HDD 115, the search expression generation unit 202B generates the above expressions (21) and (22). The search unit 203 searches for content according to the search formula generated by the search formula generation unit 202B, and the search history update unit 209 stores the search result of the search unit 203 in the HDD 115 and updates the search history.

  The search formula generation unit 202B generates a search formula for all combinations of condition elements for the updated content, and the search history update unit 209 stores the search result in the search unit 203 and updates the search history. The history creation process ends (Yes in step S304). The search operation is the same as that of the second embodiment described with reference to FIG.

  The search history of the updated content may be created at the search processing timing. In this case, the updated content is extracted in advance, and when the updated content is the target of the search formula, a new search is performed from the HDD 115 without using the search history. Alternatively, a search result for the updated content can be newly added after using the search history.

  In addition, if the updated content is not the target of the search expression (for example, if it is clear that the updated content is not the target of the search expression from the unupdated portion of the update content), search using the search history It can be performed. Even when the search history update unit 209 performs a search including the updated content in the HDD 115, the search result is stored in the search history database.

  As described above, when the content of the HDD 101 is updated, the content search device 3 according to the third embodiment creates a search history of the updated content in advance and stores it in the HDD 115. For this reason, the content can be searched more efficiently. Other effects are the same as those of the first and second embodiments.

(Other embodiments)
Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  DESCRIPTION OF SYMBOLS 1-3 ... Content search apparatus, 101 ... HDD (content DB), 102 ... Antenna, 103 ... Reception part, 104 ... Separation part, 105 ... AV decoder, 106 ... Speaker, 107 ... Monitor, 108 ... CPU, 109 ... Memory , 110 ... receiving part, 111 ... display part, 112 ... terminal, 113 ... I / F, 114 ... user I / F, 115 ... HDD (search history DB), 201 ... search condition acquisition part, 202 ... search formula generation part , 203 ... search unit, 204 ... search result analysis unit, 205 ... relevance calculation unit, 206 ... arrangement position calculation unit, 207 ... display information acquisition unit, 208 ... search result presentation unit, search history update unit 209, search history acquisition Unit 210 (extraction history acquisition unit), content update unit 211;

Claims (6)

A first storage unit storing a plurality of contents and content information including a plurality of items corresponding to the plurality of contents;
An instruction receiving unit that receives a selection instruction to select any content from the plurality of contents;
The content information of the selected content and the content information of the content stored in the first storage unit are determined to match or not match all or a part of items, and a plurality of content is determined based on the determination result An extraction unit for extracting from the first storage unit;
For each content extracted by the extraction unit, the content information of the selected content and the content information of the content extracted by the extraction unit are determined to match or not, and based on the determination result, A relevance calculator for calculating relevance with the selected content;
An arrangement position calculation unit that calculates an arrangement position of the content extracted by the extraction unit according to the degree of association calculated by the calculation unit;
A display unit for displaying a plurality of contents extracted by the extraction unit at the arrangement position calculated by the arrangement position calculation unit;
A content search apparatus comprising: The extraction unit includes:
A selection unit for selecting all or part of the items constituting the content information;
A search unit that searches for content in the first storage unit based on the selected item;
Comprising
The content search according to claim 1, wherein the selection unit selects the item until the number of contents searched by the search unit satisfies the number of contents extracted by the extraction unit or the relevance. apparatus. The instruction receiving unit receives a setting instruction of at least one of the number of contents extracted by the extraction unit or the relevance level,
The content search apparatus according to claim 2, wherein the extraction unit extracts content from the first storage unit based on the setting instruction received by the instruction reception unit. A second storage unit storing a content extraction history by the extraction unit;
An update unit for updating the content stored in the first storage unit;
The extraction unit includes:
Based on the content extraction history stored in the second storage unit, the plurality of contents are extracted from the first storage unit,
When the content is updated by the update unit,
The content information of the updated content and the content information of the content stored in the first storage unit are determined to match or not match some items, and a plurality of contents are determined based on the determination result The content search apparatus according to claim 3, wherein extraction is performed from the first storage unit, and the extraction history is stored in the second storage unit. Selecting any content from a first storage unit in which a plurality of contents and content information including a plurality of items corresponding to the plurality of contents are stored;
The content information of the selected content and the content information of the content stored in the first storage unit are determined to match or not match all or a part of items, and a plurality of content is determined based on the determination result Extracting from the first storage unit;
For each of the extracted contents, the content information of the selected content and the content information of the content extracted by the extraction unit are determined to match or mismatch, and the selected content is determined based on the determination result. Calculating the relevance to the content;
Calculating an arrangement position of the content extracted by the extraction unit according to the calculated relevance;
Displaying a plurality of contents extracted by the extraction unit at the arrangement position calculated by the arrangement position calculation unit;
A content search method comprising: Computer
A first storage unit storing a plurality of contents and content information including a plurality of items corresponding to the plurality of contents;
An instruction receiving unit that receives a selection instruction to select any content from the plurality of contents;
The content information of the selected content and the content information of the content stored in the first storage unit are determined to match or not match all or a part of items, and a plurality of content is determined based on the determination result An extraction unit for extracting from the first storage unit;
For each content extracted by the extraction unit, the content information of the selected content and the content information of the content extracted by the extraction unit are determined to match or not, and based on the determination result, A relevance calculator for calculating relevance with the selected content;
An arrangement position calculation unit that calculates an arrangement position of the content extracted by the extraction unit according to the degree of association calculated by the calculation unit;
A display unit for displaying a plurality of contents extracted by the extraction unit at the arrangement position calculated by the arrangement position calculation unit;
A content search program characterized in that the program is operated as described above.

Images