Method and apparatus for the classification of television signals

• US 5,812,732 A
• Filed: 03/22/1996
• Issued: 09/22/1998
• Est. Priority Date: 03/30/1995
• Status: Expired due to Term

• Current Assignee: Deutsche Thomson-Brandt Gmbh (Vantiva SA)
• Original Assignee: Deutsche Thomson-Brandt Gmbh (Vantiva SA)
• Inventors: Reusch, Bernd, Dettmer, Thomas, Wittner, Michael
• Primary Examiner(s): Chevalier, Robert
• Application Number:
  US08/620,870
• Time in Patent Office: 914 Days
• Field of Search: 348/5, 348/907, 348/460, 348/461, 348/468, 348/478, 348/465, 348/473, 348/474, 386/46, 386/48, 386/83
• US Class Current: 386/250
• CPC Class Codes: G11B 2220/20 Disc-shaped record carriers
• CPC Class Codes: G11B 2220/65 wherein solid state memory is used for storing indexing information or metadata
• CPC Class Codes: G11B 2220/90 Tape-like record carriers
• CPC Class Codes: G11B 27/024 on tapes G11B27/028, G11B27/029 take precedence
• CPC Class Codes: G11B 27/028 with computer assistance
• CPC Class Codes: G11B 27/107 of operating tapes
• CPC Class Codes: G11B 27/11 by using information not detectable on the record carrier
• CPC Class Codes: G11B 27/28 by using information signals recorded by the same method as the main recording G11B27/22 takes precedence
• CPC Class Codes: G11B 27/328 on a tape [TTOC]
• CPC Class Codes: H04H 60/375 Commercial
• 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 5/782 on tape
• CPC Class Codes: H04N 7/002 Special television systems not provided for by H04N7/007 - H04N7/18 still pictures via a television channel H04N1/00098
• CPC Class Codes: H04N 7/08 Systems for the simultaneous or sequential transmission of more than one television signal, e.g. additional information signals, the signals occupying wholly or partially the same frequency band , e.g. by time division H04N7/007 takes precedence

Abstract

The invention relates to the classification of such parts of television signals which are commercials so that those parts can be eliminated during recording or playback. Known methods for the detection of commercials during a TV program suffer from the drawbacks that one or two features of the TV signals are observed to decide whether the TV signal is part of a commercial or not. In the present invention of a method and apparatus, a classification of TV signals results on a plurality of information gathered from the TV signals and a plurality of rules using a rule system. Such a rule system permits the modeling of uncertainties so that for the definition of the rules, no strict values or strict limits need be given. In particular, fuzzy rules and linguistic variables are used.

Claims

1. Method for the classification of television signals, characterized by the following steps:

Defining a set of detectable features, each being characteristic of at least one property of said television signals;

Assigning a variable with a predetermined domain of arguments to each detectable feature;

Defining a predetermined functional domain for each detectable feature on said domain of arguments;

Defining a set of rule using said variables, wherein the evaluation of said set of rules results in a classification of said television signals;

Measuring said variables in said television signals for a predetermined time, and

evaluating said rules using the measurement values of said variables to classify the content of the television signals.

2. Method according to claim 1, wherein the classification results are used for the detection of commercials in TV signals.

3. Method according to claim 1, wherein the classification is performed on a framewise basis.

4. Method according to claim 1 wherein said detectable features are derived from the audio and the video parts of the television signals.

5. Method according to claim 4, wherein the following audio features are used for the definition of the variables:

stereotone and two channel audio.

6. Method according to claim 5, wherein the stereotone feature is subdivided into

stereotone indication via pilot signal

stereotone indication via VPS signal, and

stereotone indication via level difference.

7. Method according to claim 5, wherein the two channel audio is subdivided into

two channel indication via pilot signal

two channel indication via VPS signal, and

two channel indication via level difference.

8. Method according to claim 4, wherein the following features of the video signals can be analyzed with image processing:

distribution of brightness in one frame,

color variance in one frame,

measure of the difference of two successive frames,

use of VPS information

use of VPS break signal, and

use of VPS code.

9. Method according to claim 1 wherein said variables used for the classification are linguistic variables.

10. Method according to claim 1 wherein the set of variables is a fuzzy set, and fuzzy logic is used in the evaluation of the set of rules.

11. Method according to claim 10, wherein the functional domain is formed by the interval 0,1!.

12. Method according to claim 11, wherein the functions defined on the domain of arguments to achieve a fast evaluation of said set of rules are trapezoidal functions including triangles and rectangles.

13. Method according to claim 12, wherein said trapezoidal functions are characterized by four values x₁, . . . , x₄, wherein x₁ being element of the domain of arguments and z being element of a 0,1!, wherein z=T(x) is computed as follows:

space="preserve" listing-type="equation">z=0;

x μ

x.sub.1 or x.sub.4,

space="preserve" listing-type="equation">z=1;

x·

x.sub.2 and x μ

x.sub.3.

14. Method according to claim 1 wherein each rule out of the set of rules can be written in the following form:

If (condition) then variable_-- x;

feature_-- y, wherein a condition is of the form;

(not! variable_-- a is feature_-- b {and°

or} condition!.

15. Method according to claim 14, wherein the operators used in the rules are defined as follows:

a and b;

=minimum (a,b)

a or b;

=maximum (a,b)

not a;

=1-a.

16. Method according to claim 12 wherein the area and the momentum of the above functions are multiplied by the fuzzy value of the condition, wherein the area and the momentum of the function defined by x₁, . . . x₄ is computed as follows:

space="preserve" listing-type="equation">area;

=(x.sub.2 -x.sub.1)/2+(x.sub.3 =x.sub.2)+(x.sub.4 -x.sub.3)/2

space="preserve" listing-type="equation">momentum;

=(x.sub.2 -x.sub.1)/2*(x.sub.1 +(x.sub.2 -x.sub.1)* 2/3) +(x.sub.3.sup.2 -x.sub.2.sup.2 /2 +(x.sub.4 -x.sub.3)/2*(x.sub.3 +(x.sub.4 -x.sub.3)/3).

17. Method according to claim 1, wherein the classification is evaluated using the knowledge of present and past television signals, i.e. classification is performed in real time.

18. Method according to claim 17, wherein the knowledge of future television signals is also used for the evaluation of the set of rules, i.e., the classification is done using a recorded television signal.

19. Method according to claim 14 wherein the classification results are stored in a memory for future use.

20. Method according to claim 18, wherein a threshold value s is defined so that if the probability of the occurrence of a commercial is smaller than s, the data of the last 60 s are stored, and if the probability of the occurrence of the commercial is greater or equal to the threshold, the data are continuously stored until the occurrence of the commercials is smaller than the threshold for 90 s.

21. Method according to claim 20, wherein the values of the color variance, brightness variance, audio level and logogram similarity are compared at the start and the end of a data block.

22. Method according to claim 20, wherein the time intervals between the black images are detected in a data block of the television signals for the evaluation of the start and the end of a commercial.