Method and system for re-identifying broadcast segments using statistical profiles

• US 7,064,796 B2
• Filed: 12/05/2002
• Issued: 06/20/2006
• Est. Priority Date: 12/21/2001
• Status: Expired due to Fees

• Current Assignee: Eloda Corp. (Cision Ltd.)
• Original Assignee: Eloda, Inc. (Cision Ltd.)
• Inventors: Roy, Yves, Dupuis, Jean-Charles, Pouliot, Jean-François
• Primary Examiner(s): An, Shawn S.
• Application Number:
  US10/310,190
• Publication Number:
  US 20030121046A1
• Time in Patent Office: 1,293 Days
• Field of Search: 348/700, 348/462, 348/461, 348/473, 725/22, 725/18, 725/19, 725/20
• US Class Current: 348/700
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: H04H 2201/90 characterised by the use of signatures
• CPC Class Codes: H04H 60/375 Commercial
• 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
• CPC Class Codes: H04N 21/4331 Caching operations, e.g. of an advertisement for later insertion during playback
• CPC Class Codes: H04N 21/439 Processing of audio elementary streams
• CPC Class Codes: H04N 21/44008 involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59
• CPC Class Codes: H04N 21/4542 Blocking scenes or portions of the received content, e.g. censoring scenes
• CPC Class Codes: H04N 21/812 involving advertisement data advertising per se G06Q30/02

Abstract

A method and system for extracting segments from a broadcast television signal uses a mean display attribute value of pixels of each successive frame of a video component of the signal to identify segment transitions. Elapsed times between the identified segment transitions are compared with expected durations of segments, in order to detect intervals that are likely to constitute segments. Groups of the detected intervals are subjected to tests to see if they are temporally arranged in an expected distribution typical of segments in broadcast signals. The mean display attribute values concurrent with each detected segment are then used to form a statistical profile for the segment. These statistical profiles may be used to identify re-broadcast segments. Further a stream of mean audio amplitudes is preferably used to supply further segment transitions, to improve segment detection, and to provide more statistical values.

Claims

1. A method for generating a statistical profile of a segment of successive frames of a broadcast segment for re-identifying the broadcast segment, comprising steps of:

calculating at least an approximate mean value of a display attribute of each frame in a predefined set of the frames of the segment;

binning the mean values of the frames, to derive at least one attribute distribution; and

applying a statistical analysis to each of the at least one attribute distribution in order to produce a plurality of values of statistical properties that characterize the segment to be used to form the statistical profile for re-identifying the broadcast segment,

wherein the step of calculating comprises;

computing a first mean value of the attribute distributions;

computing a mean measure of deviation from a second mean value of the attribute distributions;

averaging the values of the respective display attributes of a predefined set of pixels of each frame in the predefined set of the frames, the predefined set of pixels being selected to be representative of all of the pixels in the frame;

forming respective attribute distributions from respective subsets of the predefined set of frames in the segment;

calculating the value of at least one statistical property for each of the respective attribute distributions, the each of the attribute distributions being indexed by a label of the subset from which it is formed;

parsing the mean values of the frames generated by substantially all of the frames in the segment, into respective attribute distributions, each attribute distribution being generated from a sequential set of the frames in the segment, wherein the sequential sets used to generate the attribute distributions jointly include substantially all of the frames of the segment, and a sequence order of a particular sequential set in the segment indexes the attribute distribution generated with the particular sequential set; and

multiply parsing the stream of the mean values to form a hierarchy of indexed attribute distributions wherein each particular attribute distribution is uniquely indexed by two whole numbers (n,r) where n is a level in the hierarchy, and r the sequence order of the sequential subset of frames used to generate the particular attribute distribution in an n^th level parsing of the stream.

2. A method as claimed in claim 1 wherein the step of applying comprises a step of calculating values of the statistical properties that depend on the entire attribute distribution and are substantially immune to noise and other signal transmission and reception artifacts that may be present in the segment.

3. A method as claimed in claim 1 wherein the step of computing the first mean value comprises a step of determining a median value of the attribute distributions.

4. A method as claimed in claim 1 wherein the step of computing the mean measure of deviation from the second mean value comprises a step of calculating a standard deviation of the attribute distributions.

5. A method as claimed in claim 1 wherein the step of averaging comprises steps of:

determining a value of the display attribute of a proportion of pixels coincident with one of a plurality of diagonals of the frame; and

averaging a value of the display attribute of the proportion of the pixels.

6. A method as claimed in claim 5 wherein the step of determining the value comprises a step of measuring at least a one-byte luminance value of the proportion of the pixels.

7. A method as claimed in claim 1 wherein the segment further comprises an audio component, and the method further comprises steps of:

calculating a mean audio amplitude value of the audio amplitude of each frame in the predefined set of the frames in the segment;

tallying a number of the frames in each of the subsets of the frames that have a same mean audio amplitude value, for each mean audio amplitude value, to derive respective audio distributions for each of the respective subsets; and

calculating the values of the statistical properties of each of the audio distributions to be included in the statistical profile.

8. A method as claimed in claim 7 wherein the step of calculating a mean audio value comprises a step of calculating an absolute value of a normalized amplitude of the audio component, averaged over each frame.

9. A method as claimed in claim 8 wherein the step of calculating an absolute value comprises steps of:

sampling the absolute value of the normalized amplitude of the audio component; and

averaging the samples concurrent with the frame.