Method and system for automatic detection of content

• US 9,430,472 B2
• Filed: 05/13/2013
• Issued: 08/30/2016
• Est. Priority Date: 02/26/2004
• Status: Active Grant

• Current Assignee: Mobile Research Labs, Ltd.
• Original Assignee: Mobile Research Labs, Ltd.
• Inventors: Cheung, Kwan
• Primary Examiner(s): Yaary, Michael D
• Application Number:
  US13/892,912
• Publication Number:
  US 20130318096A1
• Time in Patent Office: 1,205 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/41 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/435 Filtering based on additional data, e.g. user or group profiles
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: G10L 25/00 Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 muting semiconductor-based amplifiers when some special characteristics of a signal are sensed by a speech detector, e.g. sensing when no signal is present, H03G3/34
• CPC Class Codes: H04H 60/37 for identifying segments of broadcast information, e.g. scenes or extracting programme ID
• CPC Class Codes: H04H 60/56 Arrangements characterised by components specially adapted for monitoring, identification or recognition covered by groups H04H60/29-H04H60/54
• CPC Class Codes: H04H 60/58 of audio determination or detection of speech characteristics in general G10L25/00; speech recognition in general G10L15/00
• CPC Class Codes: H04H 60/59 of video recognising characters or patterns in general G06F18/00, G06V20/00

Abstract

A method and system for tracking use of audio and audiovisual works is described. Known are converted into a time series of frequency domain signatures. As the system detects unknown works transmitted or otherwise available for analysis, the unknown works are converted into a time series of frequency domain signatures and then the sequence of signatures matched in the database of known works. When a known work is found to have signatures that meet a matching test with an unknown work, a database is updated to reflect that the unknown work is an instance of the known work. The system includes a remote detector that receives unknown content and generates signatures that are transmitted to another location for where the matching is performed.

Claims

1. A system comprised of a central processing unit and a data memory for detecting whether a known content item is present in a plurality of unknown content items comprising:

a first module adapted to receive known content and to generate known signatures from such known content;

a database module comprised of computer memory operatively connected to the first receiving module adapted to store the generated known signatures in a database whereby each known signature is associated with a data value representing the identity of the known content;

a matching module operatively connected to the database module that receives signatures derived from the unknown content, said received unknown signatures of the same type as the known signatures stored in the database and further adapted to determine by use of the database module whether the unknown content is the known content; and

a remote receiving module that is adapted to receive unknown content, generate signatures from the unknown content of the same type as the signatures generated from the known content and transmit the generated unknown signatures to the matching module in order to cause the matching module to determine the identity of the unknown content;

where the database module is adapted to store the generated signatures with a corresponding time value and the remote receiving module is adapted to generate a time value for each generated unknown signature and transmit such time values to the matching module and where the matching module is adapted to determine whether there is at least one unknown signature that meets a predetermined matching test within a predetermined tolerance with at least one known signatures and, for the matching signatures, whether the time values corresponding to the unknown signatures and the matching known signatures are consistent with the unknown content being the known content.

2. The system of claim 1 where the matching module is further adapted to determine whether the time values associated with the matching signatures of the known content increase substantially monotonically in relation to the time values associated with the matching signatures of the unknown content.

3. A method executed by a digital signal processing system of determining the identity of broadcast unknown content comprising:

generating from a plurality of known content a series of signatures, each of the known signatures associated with a time value and storing the generated known signatures in a database and associating such stored known signatures in the database with an identifier representing the identity of the known content;

receiving from a remote device at least one unknown signature and corresponding time values where the unknown signatures are generated from unknown content and the unknown signatures are the same type of signature as that of the known content; and

determining the identity of the unknown content by determining whether there is at least one known signature in the database that meets a predetermined matching test within a predetermined tolerance with the received at least one unknown signatures,

where the determining step is further comprised of converting each of the known and unknown signatures into a numeric value and determining for each of the known signatures, which known signature has a numeric value within a tolerance value from the integer value calculated from the unknown signature.

4. The method of claim 3 where the determining step is further comprised of:

calculating an n-dimensional distance between the known signatures and the unknown signatures, where n is the number of elements in the signatures and comparing the calculated n-dimensional distance to a predetermined threshold value where the elements of the known and unknown signatures each represent unique frequency bands of sufficient width that the known and unknown signatures are not appreciably perturbed by variations in playback speed of either the known or unknown content within a predetermined playback speed tolerance value.

5. The method of claim 4 where the n-dimensional distance is one of the group:

(i) the approximate vector distance from a known signature to an unknown signature;

(ii) the approximate L-1 norm between a known signature and an unknown signature;

(iii) the approximate maximum difference between any element in the known signature and its corresponding ordinal element in any unknown signature;

(iv) the approximate minimum difference between any element in the known signature and its corresponding ordinal element in any unknown signature;

(v) the approximate average difference between all of the elements in the known signature and their corresponding elements in any unknown second signature.

6. The method of claim 3 where the predetermined matching criteria test is further comprised of:

calculating a set of absolute values of the differences between each ordinal element of the known signature and the corresponding ordinal element of any unknown signature;

calculating a sum of the set of absolute values; and

determining whether the calculated sum is less than a pre-determined value.

7. The method of claim 3 where the determining step further comprises:

determining whether the time values associated with the matching signatures of the known content increase substantially monotonically in relation to the time values associated with the matching signatures of any unknown content.

8. A system of computers adapted to perform the steps of claim 3.

9. A system of computers adapted to perform the steps of claim 4.

10. A system of computers adapted to perform the steps of claim 5.

11. A system of computers adapted to perform the steps of claim 6.

12. A system of computers adapted to perform the steps of claim 7.