CN101069180A - Adaptation of time similarity threshold in associative content retrieval - Google Patents

Adaptation of time similarity threshold in associative content retrieval Download PDF

Info

Publication number
CN101069180A
CN101069180A CNA2005800413268A CN200580041326A CN101069180A CN 101069180 A CN101069180 A CN 101069180A CN A2005800413268 A CNA2005800413268 A CN A2005800413268A CN 200580041326 A CN200580041326 A CN 200580041326A CN 101069180 A CN101069180 A CN 101069180A
Authority
CN
China
Prior art keywords
time
candidate
content item
distance
threshold
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CNA2005800413268A
Other languages
Chinese (zh)
Inventor
E·M·A·迪德里克斯
B·M·范德斯卢伊斯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Koninklijke Philips NV
Original Assignee
Koninklijke Philips Electronics NV
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Koninklijke Philips Electronics NV filed Critical Koninklijke Philips Electronics NV
Publication of CN101069180A publication Critical patent/CN101069180A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/93Document management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

Retrieval of similar content item or documents is provided based on similarity of an associated time, such as time of creation or usage. A time of one or more identified anchor document(s), in a database for example, is determined by extracting time metadata. Then a first threshold is set based on a criterion distance in time determined between the base time or current time and the anchor item time. A candidate document item time is identified and the time metadata for candidate content item may be extracted. A distance in time between the base time and the candidate time may be determined as a candidate distance in time. A candidate content item may be selected as similar for retrieval based on the first candidate distance in time and the first threshold. The notion of criterion distance in time in time- determined granularity for setting a threshold is provided.

Description

The adjustment of the time similarity threshold in the associative content retrieval
The present invention relates to content retrieval, management and oblatio field and determining based on the similarity threshold of the content item of service time and metadata.
In recent years, the storage capacity of memory device and database comprises the storage capacity of the memory device of the hard disk of personal computer and other type increasing rapidly always.Estimate that according to some storage capacity is the speed increment to double every year approximately, meanwhile, the growth of the network bandwidth is also very fast.As a result, memory device, stores a large amount of contents, need provide convenience so that memory device is carried out access to the user.The content that the user can be stored in equipment, database even the individual hardware device is flooded, unless these contents are managed in some way and organized for user's access and provide convenience, otherwise the user may not have ability to go to retrieve the content that can obtain on the network such as the internet.Do not have to adopt to the mode index of user transparent or the content of tissue, with regard to the user needs, may be " losing ", and unlikely be retrieved.
A lot of retrieval schemes are as you know arranged at present.People's such as Farnham U.S. Patent Application Publication instructions NO.2003/0158855 discloses the association of automated content, and wherein, correlativity dynamically produces between target and metadata, so that determine between each computer documents the similarity with numeric representation.People's such as Stubler U.S. Patent Application Publication instructions NO.2003/022541 discloses by produce title or the semantic marker based on the acquired image of image that obtains and the similarity between the memory image for the image extraction metadata that obtains.The U.S. Patent Application Publication instructions NO.2003/022541 of Platt discloses a kind of automatic playlist maker, and wherein several first seed-song comprise undesired seed-song, are used to generate the song in the playlist.The U.S. Patent Application Publication instructions NO.5616876 of Cluts discloses by the editor and has selected the song that similarly adds with the first suite of song song for the type mark of every first song writing in advance.The U.S. Patent Application Publication instructions NO.2004/5616876 of Gargi discloses the automatic propagation that file metadata comprises Time Created.But neither one is open to select or to refuse the setting of the mode of destination item to the threshold value of time similarity in these lists of references.The U.S. Patent Application Publication instructions NO.2002/0099696 of Prince discloses fuzzy database retrieval, selects the project that is retrieved therein to mark of similarity, and with threshold value.But, be all not open or suggestion of Prince or other reference paper based on one reference time setting threshold, also not according to reference time and the distance that is identified between the PROJECT TIME coming setting threshold.
The user certainly may the retrieval of content project, but attempting to find similar project is the thing that wastes time and energy very much, particularly when not knowing the content type of desired project or user when not specifying the content type of desired project.Further, along with content item constantly accumulation in the database of memory device or user's control, the retrieval of content item also can be more and more difficult.
A kind of method, system, unit and computer-readable medium of implementing or carrying out the various functions of searching system is provided.Selected candidate content item or a plurality of project is provided.Determined reference time.For example, may be the current time this reference time.Discern first time by the time data that extracts first content identified project.First threshold value may be set according to the standard time distance of determining between reference time and first time.Can discern candidate's time and extract candidate content item purpose time data.Reference time and the distance of first candidate between the time can be confirmed as candidate distance.It is similar that candidate content item can be selected as with first threshold value based on first candidate's time gap, being used for data base organization or retrieval, and correspondingly is that signal is selected in this of alternating content of choosing output.
The thought of the standard time distance of the form with granularity that the time determines that is used for setting threshold is provided, has come setting threshold according to it so that apart from the nearer time of reference time apart from size ratio apart from reference time far away time higher apart from granularity.
In addition, can set second threshold value based on the standard time distance, first threshold and this second threshold value have been formed a scope, if first candidate's time gap within this scope, candidate content item is just selected so.
Each very first time may comprise last time or the most frequently used time of content item of using of time, content item that content item obtains.This time may be the up-to-date correction time of reference time, content item of content item or the creation-time of content item.
The additional project of content identified of other can be identified, and time group and distance can be determined, so that first threshold also may be set based on these standard time that is determined distances.
Fig. 1 is the searching system synoptic diagram according to one embodiment of the invention.
Fig. 2 is the system operation process flow diagram according to one embodiment of the invention.
Following discussion and above-mentioned accompanying drawing have been described applicant's working of an invention scheme, this embodiment is the current best understanding of inventor, but the present invention also may much revise, embody or realize with other form under the situation that does not break away from inventive concept with other form.In addition, describe in the embodiment that some parts may be omitted under the situation that does not break away from spirit of the present invention, combination or combine with other embodiment selectively with the form of integral body, perhaps be used for replacing the feature of other embodiment or a part wherein.Therefore, accompanying drawing and detailed explanation are interpreted as illustrating various aspects of the present invention, rather than limit the scope of the invention.
As shown in Figure 1, searching system 1-1 comprises several modules, hereinafter will do being described in detail.The module of searching system 1-1, a part wherein, with and/or the integral body of searching system, may comprise hardware, software firmware or the former combination, but certain module may for example be made up of hardware, and other module may be realized by software, firmware or their combination.
Should be appreciated that the module of searching system and not requiring all is arranged in or is integrated in same equipment.Also can imagine distributed frame for searching system, this searching system can " be carried " on the appropriate module that existing device provides and be realized.
Following description will be referred to one physically by wired or link wireless link and with database 1-2 in conjunction with or the searching system 1-1 that links to each other.A clock (not shown) also may or link to each other with searching system 1-1 combination.Database 1-2 may be implemented in the memory device, for example the hard disk of personal computer, personal video recorder, entertainment systems, electronics hand account, personal hand-held equipment, Jaz disk unit or to implement such as commercial memory devices such as disc drivers.Database 1-2 is interpreted as comprising continuous several memory devices, so that the management of the content item of two or more such equipment or classification become possibility.Database may be interpreted as and comprise one or more mediums, and for example disk comprises CD, DVD, Zip floppy disk, floppy disk and data tape etc., and they can the system of being retrieved be written into and retrieve.But searching system is construed as also can come retrieval of content by the network 1-9 such as LAN (Local Area Network), wide area network and internet.
As shown in Figure 1, searching system 1-1 comprises time data extraction apparatus 1-11, and it is a module of collecting specific type of data from content item.Content item may be the multi-medium data of video, video clipping, film, photo, text, music data, audio file or other type, as jpeg file or XML data.For example, video may be one section home videos shooting on the digital VTR, film may be the cinematic data of commercial distribution, for example by the film of MPEG (comprising MPEG-2, MPEG-3 etc.) coding, photo may be digital photos data, one group of photo or photograph album, text may be file, electrical form or the computer code file that word processor generates, and music data may be the file of MP3 and so on, and so on.
Data of description by time data extraction apparatus 1-11 extracts comprises the information such as the use data of metadata or related content project.These information also comprise the time data that is used for content item, for example the acquisition time of the creation-time of project, project; At last/first/time of inferior playback second from the bottom and/or content of edit project, and time of normal use, for example, project is everlasting about 8 of afternoons most, or in the week, one month or 1 year specific one day is used, and the time that project was everlasting most such as evening is used." the most normal " used herein may be based on mean value, the intermediate value of service time, the mode of service time of service time etc.These use historical data to be sometimes referred to as metadata, and on the contrary, the type of metadata is called the use historical data sometimes.
Temporal information discussed here may be the temporal information of one or more this similarity scopes, and perhaps it may be unique or the temporal information of that scope of weights maximum.Weighing degree (if any) that these factors are reached in weight depends on and uses and user's needs.
It should be understood that, this data of description about project may search and extract in a different manner, comprise from project, from index or from the file of data base administration, or from external information sources, for example, the WWW from being connected with searching system 1-11 to the Internet 1-9 via wired or wireless link.
The content item that is identified may be discerned by one or more modes.The user may specify a project, and other project (referring to candidate content item sometimes) will be retrieved based on this project.Alternatively, adding newly or the content item that produces may be automatically to be appointed as a content item that is identified, and other project can be retrieved based on this project.
Reference time by reference time determiner 1-13 determine.Such reference time can be by user input or set, in advance programmed settings, that determine according to the clock (not shown), from the internet or current time of determining of other network array configuration that determine or in the above described manner.It should be understood that each reference time and the time relevant with candidate content item, all may comprise a date and/or time with the content identified project.Use for some, the not free date incites somebody to action enough or is even more relevant.For a lot of application, time and date is used simultaneously.It should be understood that these date and time informations and temporal information can be converted into the form that helps computing time distance and help comparing with other date and time.
Time data extraction apparatus 1-11 determines and content identified project or relevant time of project cluster, and determines time relevant with identification content item (group) and the time gap (time value that has promptly disappeared) of reference time.This distance is called the first standard time distance sometimes.The time possibility relevant with content identified project (group) is definite according to the metadata of being correlated with content item, definite according to database index, basis comprises that the network 1-9 such as WWW determines, determines by asking the user to import combination definite and in the above described manner.Time gap may be by tabling look-up, calculate, ask the user to import and the combination of aforesaid way is determined.
Threshold setting device 1-14 sets a threshold value or a scope, and the candidate content item purpose is selected to satisfy this threshold value or scope.This threshold value or scope are set according to first standard time distance by threshold setting device 1-14.
Candidate content item recognizer 1-12 connects by network or is similarly other information sources and/or based on their distance of the time gap from reference time to first content item that is identified to reference time from metadata or out of Memory for them, thereby discerns candidate content item in database.
Controller 1-15 coordinates the repertoire of searching system 1-1, carries out alternately with user interface 1-1, database 1-2, server 1-9 and the external world generally speaking, and concurrency processing system is set.
Selector switch 1-16 selects titular candidate content item, and follower 1-17 is provided for selecteed and/or unaccepted candidate content item purpose consequential signal as a result.Follower 1-17 and miscellaneous equipment carry out interface as a result, and communicate with the external world, comprise with user's (not shown) carrying out interface.Specifically, result for retrieval follower 1-17 sends signal to the content item that is retrieved by searching system 1-1 to user interface.User interface 1-3 may be a kind of separate equipment or with the miscellaneous equipment or the system integration, for example, personal computer, individual video camera, one or more storer and the above-mentioned equipment of enumerating.
Here has the available content item that is identified in addition.Therefore, tempon data and/or this process that use is extracted and time gap is determined can repeat 1 to the N any amount of available content item that is identified, and N is the positive integer greater than 1.Then, candidate content item will be selected according to the mean value of all standard time distances.
The operation of one embodiment of the invention is described below in conjunction with Fig. 1-2.As mentioned above, the first content project by user shown in Figure 1 via user interface 1-3 identification, by such as detect among the database 1-2 initiate or independently content item etc. discern automatically by system.
In S1, reference time, determiner 1-13 determined reference time by mode as mentioned above.As described in the S2 among Fig. 2, the time data extraction apparatus 1-11 of searching system 1-1 is the first content identified item extraction very first time data.In S2, other content identified project may similarly be handled (for top n content identified item extraction time data), for example, if user or system specify several " anchor point " document (coming the searched targets document based on the document).
In S3, by the time gap of decision between the time of reference time and content identified project, 1-11 determines a standard time distance for each content identified project by the tempon data extractor.In S4, may be averaged these standard time distances and obtain the average time gap.Here used be averaged to be based on calculate mathematical mean, mode or an intermediate value and draw.In addition, also can adopt the simple summation of numerical value and the aggregate concept that other statistical function of similarly suitably selecting provides selected data.
In S5, set a threshold value or a scope based on standard time distance or its mean value.For example, can specify a threshold value make value 1 or 0 may representation reference time and first content identified project between very little time gap, and numerical value 9 or 10 may be represented big time gap.Selectively, except using 1 scale between 10, threshold value may be represented such as marks such as " identical time ", " very approaching time ", " approaching time ", " remote time ", " far-out times ".It should be understood that and to use many other to be used for the scheme of these numerical value and not break away from spirit of the present invention.
In addition, second threshold value also may be selected in the same way.Therefore, first threshold may be represented the maximum time distance, and second threshold value may be represented the minimum time distance, thereby these two threshold values have constituted a scope together.The distance (reference time and the distance of candidate content item between the time) of having only candidates is within this scope the time, and candidates is just selected.
In S6, the candidates content identifier is identified in candidate content item among the database 1-2 Anywhere by network or other among Fig. 1, and time data extraction apparatus 1-11 (Fig. 2) is each candidate content item extraction time data.Then, candidate content item purpose time gap fixes among the S7 really and finishes.Other content identified project also may be available, for candidate content item 1-M, will continue the step of execution extraction time data and definite time gap numerical value.
In S8, selector switch 1-16 compares candidate content item purpose time gap and threshold value or scope.If this distance is littler or within this scope than threshold value, then select this candidate content item.
For example, if be defined as point in the mornings 8 on November 22nd, 2004 reference time, the first content identified PROJECT TIME is defined as point in the mornings 6 on November 22nd, 2004, two hours time gap so, promptly representing the distance between the reference time and the first content identified project, is the first standard time distance.Based on this time gap of two hours or based on the mean value of the standard time distance of determining by content identified project 1-N, the threshold time distance is set, for example, " 4 hours ", " on the same day ", " among one day identical time period ", " approaching time gap " or " 4 " (4 is the integers between 0 to 9, wherein 0 express time is basic identical, and 9 express times distance is very big).Then, candidate content item purpose distance compares with threshold value, in S8, if candidates time and reference time distance 4 hours with interior, within " on the same day ", within " among one day identical time period ", within " approaching time gap " or within the time gap of grade " 4 " threshold value, this candidates is just selected.
According to an aspect of the present invention, setting threshold so that apart from the nearer time of reference time apart from size ratio apart from the time far away reference time apart from the granularity height.Therefore, for example, if arrange by scale 1 to 10 apart from the distance of reference time, along with the increase of the distance of distance reference time, long distance will be comprised in less several scales so.Therefore, if standard time distance (document of reference time and identification or the distance between the content item) is 1 hour, if be 5 hours apart from the candidate distance of this benchmark, first candidate content item may be judged to be dissmilarity so.But if the standard time distance is 1000 days, the judgement of second candidate documents may be similar so, even second candidate is 6000 days apart from this benchmark.Such threshold value determines to be based on following thought: people are thinking usually instinctively, it is important that for a long time the difference of the time gap between Zhi Qian the example does not have the difference of the identical time gap between the nearest bygone example: the time gap from relevant reference time is far away more, is then determining that the time gap between each example is just inessential more aspect the similarity.The threshold mode of getting like this is sometimes referred to as the granularity that standard time distance determines here and gets threshold mode.
As mentioned above, on the dispersed basis of the set existence of content identified project distance value, in S5, threshold setting device 1-14 also can utilize max-thresholds and minimum threshold and produce a scope.Therefore, adopt the aforesaid example that has only a content identified project, max-thresholds is identical with above-mentioned discussion, and minimum threshold may be, for example, " different hour ", " at least one hour ", " very approaching time " or distance are rated 2.So, the scope of selecteed content item is: " on the same day different hour ", " on the same day in the same time period different hour ", " 1 to 4 hour ", " very approaching time gap-approaching time gap " or scale " 2-4 ", its concrete scope depends on the scope of system/use of deciding threshold value.
According to an aspect of invention, can use a plurality of " benchmark " times, the granularity that the standard time distance is determined will be applied to each such reference time respectively.For example, the time of the current time of a reality and one major event in the past as date of birth of wedding date of user, child, anniversary etc., may be second reference time.Therefore, the rank of granularity will reduce with the distance of distance reference time 1 (past candidate documents time remote more, the time quantum that is considered to the disappearance similar to it is just many more), and similarly, the granularity rank will reduce with the distance of distance reference time 2.The idea here is as follows: just important more such as the people who haves a wedding in the specific date for concerning the someone apart from the temporal difference that second reference time is near more, so it needs higher granularity.
Like this second, third, L can be set by the user reference times such as (L are the integer greater than 3) or by system according to determining such as the mode of above-mentioned definite reference time.Therefore, for example, if system (for example detects a large amount of or very big percentage and subscriber-related document, document resides among user's computer, database, the palmtop computer) content item time of having is (for example to be in special time, the date of creating, last service time etc.), so just can set this additional reference time.For example, a large amount of content items, they for example use a certain threshold value or show that based on statistical function its content item is more concentrated than normal condition, as an example, as the wedding photo before and after user's wedding same day or wedding, wedding video, music, Email etc., then just can be confirmed as a so additional reference time.Then, standard time distance can be determined, and on so additional reference time basis, comes setting threshold according to the granularity that the standard time distance is determined.
The content type of the content item that the content type of the content item that retrieves according to an aspect of the present invention, may be selected with the user is different.For example, if the content item that the user selects is the file or the MP3 of music type, the content item that then retrieves may be the content type photo data.With like this, for example, based on time similarity, the photo of particular type may be retrieved so that the music of selecting with the user is complementary.
In S9, these (these) selected candidate content item is provided for user or user interface 1-3.A signal may be provided directly to database 1-2, so that selected candidates is retrieved in database or user interface 1-3.(extraly) alternatively if candidate content item is rejected, then can provide a signal.
May provide a notice to notify user's (not shown) searchable content item to user interface 1-3.This notice may comprise the retrieval and the aforesaid combination of whole or its part of the sign of content item to be retrieved, the description of content item, a URL(uniform resource locator) (URL) of pointing to content item or link, content item.This system also may be used for projects combo with anchor item and selected retrieval together so that tissue database.In S10, EO.
Above-mentioned embodiment of the present invention that provide only are as illustration.But, it should be understood that scope of the present invention provides in the claims.

Claims (13)

1, a kind of method of content item retrieval comprises:
Determine (S1) reference time;
The tempon data of extracting (S2) first content identified project are with as the very first time;
Set (S5) first threshold based on a standard time distance between the reference time and the very first time, determining (S3);
Extract (S6) first candidate content item purpose tempon data with as first candidate's time, and determine that (S7) reference time and the distance of first candidate between the time are with as first candidate's time gap;
Select (S8) first candidate content item based on first candidate's time gap and first threshold; And
When first candidates is selected, be that signal is selected in one of first candidate content item output (S9).
2, the method in the claim 1, wherein first threshold is set based on the definite granularity of standard time distance.
3, the method in the claim 1 further comprises based on standard time distance setting second threshold value, and second threshold value and first threshold are formed a scope, and works as first candidate's time gap within this scope, then selects first candidate content item.
4, the method in the claim 1, wherein, the very first time and at least one in the time of first candidate comprise the last time of using of time, content item of content item acquisition, time and the most frequently used time of content item that content item uses for the first time.
5, the method in the claim 1, wherein, first time and at least one in the time of first candidate comprise the time of the up-to-date correction of content item and the creation-time of content item.
6, the method in the claim 1 further comprises:
The tempon data of extracting (S2) second content identified project are with as second time;
Also set (S5) first threshold based on second standard time distance, this second standard time distance is confirmed as the distance between the reference time and second time.
7, the searching system of a content item comprises:
Extraction apparatus reference time (1-13) is configured to determine reference time;
Tempon data extractor (1-11), the tempon data that are configured to the first content identified project of extracting be with as the very first time, and the time gap between the definite reference time and the very first time is with as the standard time distance;
Threshold setting device (1-14) is configured to based on as distance between the reference time and the very first time and definite standard time distance is set first threshold;
Above-mentioned meta-data extractor (1-11) is configured to extract the first candidate content item purpose metadata with as first candidate's time, and definite reference time and the time gap of first candidate between the time are with as first candidate's time gap;
Selector switch (1-16) is configured to select first candidate content item based on first candidate's time gap and first threshold;
Follower (1-17) as a result, being configured in first candidate's time when selected is that signal is selected in one of first candidate content item output.
8, the system in the claim 7, wherein first threshold value is set based on the granularity of being determined by the standard time distance.
9, the system in the claim 7 comprises that further above-mentioned threshold setting device (1-14) is based on standard time distance setting second threshold value, second threshold value and first threshold are formed a scope, when first candidate's time gap within this scope, then above-mentioned selector switch is selected first candidate content item.
10, the system in the claim 7, wherein, the very first time and at least one in the time of first candidate comprise the last time of using of time, content item of content item acquisition, time and the most frequently used time of content item that content item uses for the first time.
11, the system in the claim 7, wherein, the very first time and first candidate comprise one of at least the time of the up-to-date correction of content item and the creation-time of content item in the time.
12, the system in the claim 7, wherein, time period, user's most recently used time period and the user that the very first time and at least one in the time of first candidate comprise frequent use be one of time period of normal use.
13, the system in the claim 7 comprises:
The tempon data that above-mentioned tempon data extractor (1-11) is configured to the second content identified project of extracting are with as second time;
Above-mentioned threshold setting device (1-14) is configured to also set first threshold based on second standard time distance that is determined as the distance between the reference time and second time.
CNA2005800413268A 2004-12-01 2005-11-30 Adaptation of time similarity threshold in associative content retrieval Pending CN101069180A (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US63213604P 2004-12-01 2004-12-01
US60/632,136 2004-12-01

Publications (1)

Publication Number Publication Date
CN101069180A true CN101069180A (en) 2007-11-07

Family

ID=36169210

Family Applications (1)

Application Number Title Priority Date Filing Date
CNA2005800413268A Pending CN101069180A (en) 2004-12-01 2005-11-30 Adaptation of time similarity threshold in associative content retrieval

Country Status (5)

Country Link
EP (1) EP1820125A1 (en)
JP (1) JP2008522309A (en)
KR (1) KR20070086805A (en)
CN (1) CN101069180A (en)
WO (1) WO2006059293A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4812031B2 (en) * 2007-03-28 2011-11-09 Kddi株式会社 Recommender system

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2002323413A1 (en) * 2001-08-27 2003-03-10 Gracenote, Inc. Playlist generation, delivery and navigation
US6987221B2 (en) * 2002-05-30 2006-01-17 Microsoft Corporation Auto playlist generation with multiple seed songs
US6996390B2 (en) * 2002-06-26 2006-02-07 Microsoft Corporation Smart car radio
US7228054B2 (en) * 2002-07-29 2007-06-05 Sigmatel, Inc. Automated playlist generation

Also Published As

Publication number Publication date
EP1820125A1 (en) 2007-08-22
JP2008522309A (en) 2008-06-26
KR20070086805A (en) 2007-08-27
WO2006059293A1 (en) 2006-06-08

Similar Documents

Publication Publication Date Title
CN100573514C (en) The adjustment of the location similarity threshold in the associative content retrieval
US7751632B2 (en) Intelligent, dynamic, long-term digital surveilance media storage system
KR100934537B1 (en) Method and apparatus for automatically detecting data type for data type dependent processing
US10515133B1 (en) Systems and methods for automatically suggesting metadata for media content
US6993535B2 (en) Business method and apparatus for employing induced multimedia classifiers based on unified representation of features reflecting disparate modalities
US20080085053A1 (en) Sampling image records from a collection based on a change metric
CN111177432B (en) Large-scale image retrieval method based on hierarchical depth hash
US20120005214A1 (en) Ordered index
CN104182413A (en) Method and system for recommending multimedia content
JP2008522311A (en) Automatic content organization based on content item association
JP4779961B2 (en) Document selection apparatus and document selection program
CN113655968A (en) Unstructured data storage method
KR20070086806A (en) Associative content retrieval
CN101069180A (en) Adaptation of time similarity threshold in associative content retrieval
JP2010218390A (en) Stream accumulation control device
KR101277419B1 (en) Data retrieval device
JP5620038B2 (en) How to automatically archive and use video that suits your purpose
CN103003880B (en) Representative image is determined for video
KR101203931B1 (en) Posting list compression and ranking method for efficient searching a person on videos
CN110888841A (en) Patent retrieval method and system based on block chain
CN117668047A (en) Method and system for improving property system filing bill query efficiency
WO2007036888A2 (en) A method and apparatus for segmenting a content item
Lee et al. Video PARIC (PArsing and Retrieval based on Image Contents)-an approach for video information retrieval system
AOKI et al. MPEG-7 Content-Based Image Retrieval with Spatial and Temporal Information

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication