Hypervideo: information retrieval using realtime buffers

• US 6,493,707 B1
• Filed: 10/29/1999
• Issued: 12/10/2002
• Est. Priority Date: 10/29/1999
• Status: Expired due to Term

• Current Assignee: Level 3 Communications LLC (Lumen Technologies, Inc.)
• Original Assignee: Verizon Laboratories Incorporated (Verizon Communications Inc.)
• Inventors: Xu, Haiping, Dey, Jayanta Kumar
• Primary Examiner(s): Corrielus, Jean M.
• Assistant Examiner(s): To, Baoquoc N
• Application Number:
  US09/431,293
• Time in Patent Office: 1,138 Days
• Field of Search: 707/3, 707/6, 707/501, 707/501.1, 707/434, 707/320, 345/719
• US Class Current: 1/1
• CPC Class Codes: G06F 16/40 of multimedia data, e.g. slideshows comprising image and additional audio data retrieval of still image data G06F16/50; retrieval of audio data G06F16/60; retrieval of video data G06F16/70
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Disclosed is a method and device for selecting documents, such as Web pages or sites, for presentation to a user, in response to a user expression of interest, during the course of presentation to the user of a document, such as a video or audio selection, whose content varies with time. The method takes advantage of information retrieval techniques to select documents related to the portion of the temporal document in which the user has expressed interest. The method generates the search query to use to select documents by reference to text associated with the portion of the temporal document in which the user has expressed interest, as by using the closed caption test associated with the video, or by using speech recognition techniques. The method further uses a weighting function to weigh the terms used in the search query, depending on their temporal relationship to the user expression of interest. The method uses a buffer to save necessary information about the temporal occurrence of terms in the temporal document.

Claims

1. A method for finding documents which relate to a portion of a temporal document, comprising:

(a) in response to a signal of interest at a particular time during the temporal document, identifying a portion of the temporal document for which related documents are to be found;

(b) selecting text associated with the portion of the temporal document identified, by reference to a buffer wherein is stored each term in text associated with the temporal document, together with a time in the temporal document at which the term occurs;

(c) weighting each term in the text selected by a function W(t) according to the time t at which the text occurs relative to the time at which the signal of interest occurs;

(d) finding the related documents by use of information retrieval techniques as applied to the text selected;

wherein each term in the text associated with the temporal document is determined as that portion of the document is accessed, and is stored in a buffer together with the time in the temporal document at which the term occurs; and

wherein $W<br><br>\left(t\right)=W<br><br>\text{\hspace{1em}<br><br>}<br><br>t_1,t_2<br><br>\left(t\right)=<br><br><br><br>{\int <br><br>}_{k=0}^t<br><br>\left(1-\exp <br><br>\left(-t_1\right)\right)*\exp <br><br>\left(-t_1<br><br>k\right)*\left(1-\exp <br><br>\left(-t_2\right)\right)*\exp <br><br>\left(-t_2<br><br>\left(t-k\right)\right),$

where t₁ and t₂ are constants.

2. The method of claim 1, wherein t₁=0.0001 and t₂=0.00025.

3. The method of claim 2, wherein W(t)=W_{t1, t2}(t) for times from 30 seconds before the signal of interest is given until the signal is given, and W(t)=0 for all other times.

4. The method of claim 1, wherein the related documents are accessed through the Internet.

5. The method of claim 4, further including selecting the related documents from among a collection of documents which may be accessed through the Internet, by utilizing databases comprising information about the collection.

6. The method of claim 5, wherein the related documents are selected from the collection according to the scores achieved when evaluating the documents in the collection according to a formula giving scores to documents depending upon the occurrence in the documents of terms which occur in text associated with the portion of the temporal document identified, where each term is weighted by a function W(t) according to the time t at which the term occurs relative to the time at which the signal of interest occurs.

7. The method of claim 6, wherein the terms which occur in the text which is associated with the portion of the temporal document identified, and the times at which the terms occur, are determined by reference to the buffer which contains the terms and the times in the video or audio material at which the terms occur.

8. The method of claim 6, where a predetermined number of documents, 1000, is selected.

11. The method of claim 6, wherein the determination of the documents in the collection which receive the highest scores is carried out using compressed document surrogates.

12. The method of claim 6, wherein the determination of the documents in the collection which receive the highest scores is carried out by a server which is distinct from the server which receives the signal of interest.

9. A method for finding documents which relate to a portion of a temporal document, comprising:

(a) in response to a signal of interest at a particular time during the temporal document, identifying a portion of the temporal document for which related documents are to be found;

(b) selecting text associated with the portion of the temporal document identified, by reference to a buffer wherein is stored each term in text associated with the temporal document, together with a time in the temporal document at which the term occurs;

(c) weighting each term in the text selected by a function W(t) according to the time t at which the text occurs relative to the time at which the signal of interest occurs;

(d) finding the related documents by use of information retrieval techniques as applied to the text selected; and

(e) selecting the related documents from among a collection of documents which may be accessed through the Internet, by utilizing databases comprising information about the collection;

wherein each term in the text associated with the temporal document is determined as that portion of the document is accessed, and is stored in a buffer together with the time in the temporal document at which the term occurs;

wherein the related documents are selected from the collection according to the scores achieved when evaluating documents in the collection according to a formula giving scores to documents depending upon the occurrence in the documents of terms which occur in text associated with the portion of the temporal document identified, where each term is weighted by a function W(t) according to the time t at which the term occurs relative to the time at which the signal of interest occurs; and

wherein a score S_D of a document D in the collection may be determined by crediting the document D, for each term T in the temporal portion of the document identified which occurs in the document D, with an amount proportional to W(t), to Robertson'"'"'s term frequency TF_TD and to ID_FT where

W(t) is the weight assigned to a term which occurs at time t relative to the signal of interest,

10. The method of claim 9, wherein K₁ is 0.5, K₂ is 1.5, K₃ is 0.5, and K₄ is 1.0.

13. A device for finding documents which relate to a portion of a temporal document, comprising:

(a) means for identifying a portion of the temporal document for which related documents are to be found, in response to a signal of interest at a particular time during the temporal document;

(b) means for selecting text associated with the portion of the temporal document identified, by reference to a buffer wherein is stored each term in text associated with the temporal document, together with a time in the temporal document at which the term occurs;

(c) means for weighting each term in the text selected by a function W(t) according to the time t at which the text occurs relative to the time at which the signal of interest occurs;

(d) means for finding the related documents by use of information retrieval techniques as applied to the text selected;

wherein each term in the text associated with the temporal document is determined as that portion of the document is accessed, and is stored in a buffer together with the time in the temporal document at which the term occurs;

wherein $W<br><br>\left(t\right)=W<br><br>\text{\hspace{1em}<br><br>}<br><br>t_1,t_2<br><br>\left(t\right)=<br><br><br><br>{\int <br><br>}_{k=0}^t<br><br>\left(1-\exp <br><br>\left(-t_1\right)\right)*\exp <br><br>\left(-t_1<br><br>k\right)*\left(1-\exp <br><br>\left(-t_2\right)\right)*\exp <br><br>\left(-t_2<br><br>\left(t-k\right)\right),$

where t₁ and t₂ are constants.

14. The device of claim 13, wherein t₁=0.0001 and t₂=0.00025.

15. The device of claim 14, wherein W(t)=W_{t1, t2}(t) for times from 30 seconds before the signal of interest is given until the signal is given, and W(t)=0 for all other times.

16. The device of claim 13, wherein the related documents are accessed through the Internet.

17. The device of claim 16, further including means for selecting the related documents from among a collection of documents which may be accessed through the Internet, by utilizing databases comprising information about the collection.

18. The device of claim 17, wherein th e related documents are selected from the collection according to the scores achieved when evaluating the documents in the collection according to a formula giving scores to documents depending upon the occurrence in the documents of terms which occur in text associated with the portion of the temporal document identified, where each term is weighted by a function W(t) according to the time t at which the term occurs relative to the time at which the signal of interest occurs.

19. The device of claim 18, wherein the terms which occur in the text which is associated with the portion of the temporal document identified, and the times at which the terms occur, are determined by reference to the buffer which contains the terms and the times in the video or audio material at which the terms occur.

20. The device of claim 18, where a predetermined number of documents, 1000, is selected.

23. The device of claim 18, wherein the determination of the documents in the collection which receive the highest scores is carried out using compressed document surrogates.

24. The device of claim 18, wherein the determination of the documents in the collection which receive the highest scores is carried out by a server which is distinct from the server which receives the signal of interest.

21. A device for finding documents which relate to a portion of a temporal document, comprising:

(a) means for identifying a portion of the temporal document for which related documents are to be found, in response to a signal of interest at a particular time during the temporal document;

(b) means for selecting text associated with the portion of the temporal document identified, by reference to a buffer wherein is stored each term in text associated with the temporal document, together with a time in the temporal document at which the term occurs;

(c) means for weighting each term in the text selected by a function W(t) according to the time t at which the text occurs relative to the time at which the signal of interest occurs;

(d) means for finding the related documents by use of information retrieval techniques as applied to the text selected;

(e) means for selecting the related documents from among a collection of documents which may be accessed through the Internet, by utilizing databases comprising information about the collection;

wherein each term in the text associated with the temporal document is determined as that portion of the document is accessed, and is stored in a buffer together with the time in the temporal document at which the term occurs;

wherein the related documents are selected from the collection according to the scores achieved when evaluating documents in the collection according to a formula giving scores to documents depending upon the occurrence in the documents of terms which occur in text associated with the portion of the temporal document identified, where each term is weighted by a function W(t) according to the time t at which the term occurs relative to the time at which the signal of interest occurs; and

wherein a score S_D of a document D in the collection may be determined by crediting the document D, for each term T in the temporal portion of the document identified which occurs in the document D, with an amount proportional to W(t), to Robertson'"'"'s term frequency TFTD and to ID_FT where

W(t) is the weight assigned to a term which occurs at time t relative to the signal of interest,

22. The device of claim 21, wherein K₁ is 0.5, K₂ is 1.5, K₃ is 0.5, and K₄ is 1.0.