Audio/Video fingerprint search accuracy using multiple search combining

• US 8,266,142 B2
• Filed: 06/04/2008
• Issued: 09/11/2012
• Est. Priority Date: 06/06/2007
• Status: Expired due to Fees

• Current Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Original Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Inventors: Jiang, Wenyu, Bauer, Claus
• Primary Examiner(s): Cao, Phuong Thao
• Application Number:
  US12/663,057
• Publication Number:
  US 20100205174A1
• Time in Patent Office: 1,560 Days
• Field of Search: 707/725, 707/914, 707/915, 707/999.003
• US Class Current: 707/725
• CPC Class Codes: G06F 16/634 Query by example, e.g. query by humming
• CPC Class Codes: G06F 16/683 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/783 using metadata automatically derived from the content
• CPC Class Codes: G06V 20/48 Matching video sequences

Abstract

A method for use in identifying a segment of audio and/or video information comprises obtaining a query fingerprint at each of a plurality of spaced-apart time locations in said segment, searching fingerprints in a database for a potential match for each such query fingerprint, obtaining a confidence level of a potential match to a found fingerprint in the database for each such query fingerprint, and combining the results of searching for potential matches, wherein each potential match result is weighted by respective confidence level. A confidence level may be a function of at least one or both of (1) a measure of difference between a query fingerprint and a found fingerprint and (2) the relative timing relationship between the time location of a query fingerprint and the time location of a found fingerprint.

Claims

1. A method for use in identifying a segment of audio and/or video information, comprising

obtaining a query fingerprint at each of a plurality of spaced-apart time locations in said segment,

searching fingerprints in a database of fingerprints for a potential match for each such query fingerprint,

obtaining a confidence level of a potential match to a found fingerprint in the database for each such query fingerprint, wherein the confidence level is a function of the relative timing relationship between the time location of a query fingerprint and the time location of a found fingerprint, and

combining the results of searching for potential matches either to identify said segment of audio and/or video information or to conclude that said segment does not belong to the database, wherein each potential match result is weighted by a respective confidence level.

2. The method according to claim 1 wherein a confidence level is also a function of a measure of difference between a query fingerprint and a found fingerprint.

3. The method according to claim 2 that comprises:

identifying said segment of audio and/or video from a search having a minimum difference in relative timing between query fingerprints and found fingerprints if the minimum difference is less than a threshold, wherein the threshold varies depending on a number of searches that identify the same segment of audio and/or video.

4. The method according to claim 2 wherein:

the relative timing relationship between the time location of query fingerprints and the time location of found fingerprints is a difference in time location, and

query fingerprints are deemed to identify a segment of audio and/or video information if the difference in time between the query fingerprints and found fingerprints is a minimum and is less than a dynamic threshold that is raised when the measure of difference between query fingerprints and found fingerprints is low.

5. The method according to claim 1 wherein said time locations are spaced apart so as to minimize correlation of the search results between or among query fingerprints.

6. The method according to claim 1 wherein there are multiple sets of spaced-apart time locations and the results of searching for each of the multiple sets are combined.

7. The method according to claim 6 wherein the results of searching are combined within each set and the results of such combinations are further combined.

8. The method according to claim 6 wherein the results of searching are combined across all sets.

9. The method according to claim 1 that comprises determining confidence of the combined results of searching using a measure of confidence of a linear trend of the relative timing relationships between the time location of the query fingerprint and the time location of the potentially matching found fingerprint at the time locations.

10. The method according to claim 1 wherein:

the results of searching are combined by a voting scheme,

the combining is cascaded by additional iterations of multiple searches in the event of inconclusive votes, and

votes are either considered only in new iterations or are combined with some or all previous votes.