Methods for Identifying Audio or Video Content

• US 20080208849A1
• Filed: 05/02/2008
• Published: 08/28/2008
• Est. Priority Date: 12/23/2005
• Status: Active Grant

• Current Assignee: Digimarc Corporation
• Original Assignee: Digimarc Corporation
• Inventors: Conwell, William Y.
• Granted Patent: US 8,341,412 B2
• Time in Patent Office: Days
• US Class Current: 707/5
• 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 16/43 Querying
• CPC Class Codes: G06F 16/489 using time information
• CPC Class Codes: G06F 16/58 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• 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: G06F 18/00 Pattern recognition
• CPC Class Codes: G06N 5/048 Fuzzy inferencing
• CPC Class Codes: G06Q 10/00 Administration; Management
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: G06Q 50/18 Legal services
• CPC Class Codes: G06Q 50/265 Personal security, identity or safety
• CPC Class Codes: H04L 63/08 for authentication of entities cryptographic mechanisms or cryptographic arrangements for entity authentication H04L9/32
• CPC Class Codes: H04L 9/3247 involving digital signatures

Abstract

The disclosed technology generally relates to methods for identifying audio and video entertainment content. Certain shortcomings of fingerprint-based content identification can be redressed through use of crowdsourcing techniques.

Claims

1. A recognition method useful in matching a first item of visual or audio entertainment content to a counterpart in a universe of second items of visual or audio entertainment content, comprising:

receiving fingerprint data derived from first visual or audio entertainment content;

identifying plural items of second content that are possible matches of said first content, based on the received fingerprint data;

sending to each of one or more human reviewers at least a portion of the first content item, and at least a portion of one of the second content items;

collecting assessment data from said one or more reviewers; and

using said assessment data in identifying one of said items of second entertainment content that appears to be a best match with said first entertainment content item.

2. The method of claim 1 that includes sending to at least one reviewer at least a portion of the first entertainment content item, and at least a portion of plural items of said second entertainment content, and collecting from said reviewer assessment data that indicates which of the plural items of sent second entertainment content appears to be a best match with the first entertainment content.

3. The method of claim 1 in which said identifying includes combining said assessment data with algorithmically determined match information to identify a match.

4. A recognition method useful in matching a first item of visual or audio entertainment content to a counterpart in a universe of second items of entertainment content, comprising:

sending to a human reviewer, over a computer network, at least a portion of the first entertainment content item;

collecting assessment data from said reviewer;

using said assessment data to prune the universe of second entertainment content items to a smaller subset; and

searching for a counterpart to the first item of entertainment content in the smaller subset of second entertainment content items.

5. The method of claim 4 wherein said assessment data serves to indicate at least one of gender, age, topical visual classification, audio type, or audio genre.

6. A method comprising:

distributing task data to plural humans over a computer network, the task data being unrelated to facial recognition;

collecting task result data from said humans, over the network; and

taking into account respective expertises of different of said humans in using said task result data.

7. The method of claim 6 that includes weighting result data from relatively more expert humans more highly than result data from relatively less expert humans, when making use of said task result data.

8. A method comprising:

computing a fingerprint from first audio or visual entertainment content;

checking a database for a match between said computed fingerprint and a collection of reference fingerprints;

in at least certain cases, referring the first entertainment content to one or more human reviewers over a computer network, for an assessment; and

storing assessment data received from the one or more human reviewers in a database, in association with said computed fingerprint, for later reference.

9. A method of processing audio or visual entertainment content to yield a fingerprint identifier different than a fingerprint identifier associated with said entertainment content in an unprocessed state, comprising at least one of the following (a)-(f):

(a) applying to said entertainment content a non-flat frequency equalization process;

(b) applying to said entertainment content a frequency equalization process that changes over time;

(c) changing the dynamic range of at least certain frequency bands of said entertainment content;

(d) altering the entertainment content in a manner exploiting temporal- or frequency-psychoacoustic masking phenomena;

(e) varying the entertainment content in a manner to exploit threshold biasing phenomena; and

/or

(f) processing the entertainment content so that it has an attribute placing it in a classification of entertainment content that is pruned from consideration as a fingerprint match;

wherein the processed entertainment content may escape recognition by automated techniques adapted to recognize the entertainment content in its unprocessed state.

10. A method comprising:

sending first and second items of data to each of one or more human reviewers, over a computer network, said items being unrelated to verifying a person'"'"'s identity;

soliciting an assessment from said one or more reviewers as to whether the first and second of items should be deemed to match, despite a difference therebetween;

receiving assessment data from said one or more human reviewers; and

using said assessment data in an automated process.