Video detection system and methods

• US 8,326,043 B2
• Filed: 04/13/2008
• Issued: 12/04/2012
• Est. Priority Date: 04/13/2007
• Status: Active Grant

• Current Assignee: IVITEC GmbH
• Original Assignee: Ipharro Media GmbH (IVITEC GmbH)
• Inventors: Cavet, Rene
• Primary Examiner(s): CHAWAN, SHEELA C
• Application Number:
  US12/595,401
• Publication Number:
  US 20100329547A1
• Time in Patent Office: 1,696 Days
• Field of Search: 382/190, 382/160, 382/162, 382/199, 382/165, 382/232, 382/166, 382/224, 382/225, 382/226, 382/229, 382/230, 382/266, 382/305, 382/306, 341/50, 341/61, 341/110, 341/126, 341/144, 341/155, 341/120, 341/133, 341/152, 341/145, 348/571, 348/655, 348/656, 348/657, 348/E9.001, 348/E9.037, 348/208.2, 348/E5.065, 702/127, 702/145, 318/638, 318/652
• US Class Current: 382/190
• CPC Class Codes: G06F 16/785 using colour or luminescence
• CPC Class Codes: G06F 16/7864 using domain-transform features, e.g. DCT or wavelet transform coefficients
• CPC Class Codes: G06V 10/435 Computation of moments
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 20/40 in video content extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44

Abstract

A video detection system and method compares a queried video segment to one or more stored video samples. Each of the queried video segments and stored video samples can be represented by respective digital image sets. A first and second comparison comprises comparing a set of low and high resolution temporal and spatial statistical moments in a COLOR9 space, and eliminating file digital image sets that do not match the queried digital image set. A third comparison generates a set of matching files by comparing a set of wavelet transform coefficients in a COLOR9 space. RGB bit-wise registration and comparison of one or more subframes of specific frames in the queried digital image set to a corresponding set of matching file subframes determines queried subframe changes. In the event of a change in a queried subframe, the changed subframe is added to the set of matching file subframes.

Claims

1. A video detection method, comprising:

(a) extracting a feature data set from a queried digital image set and a file digital image set in a first set of file digital image sets;

(b) measuring one or more errors between the feature data set extracted from the queried digital image set and the file digital image set;

(c) repeating steps (a)-(b) for one or more file digital image sets in the first set of file digital image sets;

(d) generating a second set of file digital image sets, wherein the second set consists of each of the file digital image sets in the first set having a corresponding one or more errors less than a corresponding set of thresholds;

(e) repeating steps (a)-(d) for one or more feature data sets, generating a matching set of file digital image sets;

(f) measuring one or more bitmap matching errors between one or more subframes of a queried digital image in the queried digital image set and one or more sets of one or more subframes of a file digital image in a file digital image set in the matching set of file digital image sets, wherein the bitmap comprises a set of ordered pixels in an RGB color space;

(g) translating, resizing, and requantizing the one or more subframes of the queried digital image and repeating (f) to reduce the one or more bitmap matching errors;

(h) adding, to the one or more sets of one or more subframes of the file digital image, one of;

(1) one or more new subframes, and (2) a new set of one or more subframes; and

(i) repeating (f)-(h) for one or more file digital images in one or more file digital image sets in the matching set of file digital image sets.

2. The method of claim 1, wherein the extracting in (a) comprises extracting the feature data set by a first function of one or more two-dimensional statistical moments in a COLOR9 space.

3. The method of claim 2, wherein the first function comprises:

(i) temporally segmenting the queried digital image set and the file digital image set into a set of segmented queried digital image subsets and a set of segmented file digital image subsets according to a distance measure of the one or more two-dimensional statistical moments between consecutive queried digital images in the queried digital image set and between consecutive file digital images in the file digital image set; and

(ii) finding one or more queried medians and a corresponding queried image median frame for one or more segmented queried digital image subsets from the one or more two-dimensional statistical moments, and finding one or more file medians and a corresponding file image median frame for one or more segmented file digital image subsets from the one or more two-dimensional statistical moments.

4. The method of claim 3, wherein the measuring in (b) comprises measuring one or more errors between the one or more queried medians and the one or more file medians.

5. The method of claim 3, wherein the extracting in (a) comprises:

extracting a set of two-dimensional wavelet transform coefficients in a COLOR9 space from the queried image median frame for one or more segmented queried digital image subsets; and

extracting a set of two-dimensional wavelet transform coefficients in a COLOR9 space from the file image median frame for one or more segmented file digital image subsets.

6. The method of claim 5, wherein the set of two-dimensional wavelet transform coefficients are Haar wavelet transform coefficients.

7. The method of claim 5, wherein the COLOR9 space is transformed to a grey scale.

8. The method of claim 3, wherein the queried digital image and the file digital image in (f) comprise the queried median image frame and the file median image frame, respectively.

9. The method of claim 2, wherein the set of two-dimensional statistical moments comprises at least one of:

a mean, a variance, and a skew.

10. The method of claim 1, wherein the extracting in (a) comprises extracting the feature data set by a second function of two-dimensional statistical moments in a COLOR9 space.

11. The method of claim 10, wherein the second function comprises:

(i) temporally indexing a set of queried digital images in the queried digital image set and a set of file digital images in the file digital image set; and

(ii) assigning a corresponding set of one or more two-dimensional statistical moments to the temporally indexed set of queried digital images and the temporally indexed set of file digital images.

12. The method of claim 11, wherein the measuring in (b) comprises convolving the temporally indexed set of one or more two-dimensional statistical moments for the set of queried digital images and the temporally indexed set of one or more two-dimensional statistical moments for the set of file digital images.

13. The method of claim 10, wherein the set of two-dimensional statistical moments comprises at least one of:

a mean, a variance, and a skew.

14. The method of claim 1, wherein the set of ordered pixels comprises one or more sets of:

a horizontal offset, a vertical offset, a horizontal segment length, and a vertical segment length.

15. The method of claim 14, wherein the set of ordered pixels is contiguous.

16. The method of claim 1, wherein the translating, resizing, requantizing, and repeating (f) to reduce the one or more bitmap matching errors in (g) comprises one or more feedback loops.

17. The method of claim 1, wherein the translating, resizing, requantizing, and repeating (f) to reduce the one or more bitmap matching errors in (g) comprises an edge detection and suppression function.

18. The method of claim 1, wherein the adding in (h) further comprises adding a subset of the one or more subframes of the queried digital image.

19. An apparatus adapted for producing video detection, comprising:

(a) means for extracting a plurality of feature data sets common to a queried digital image set and for each of a first set of file digital image sets;

(b) means for sequentially searching the plurality of feature data sets for a matching set of file digital image sets, wherein the sequential searching comprises matching the plurality of feature data sets common to the queried digital image set and to a sequentially smaller set of subsets of the first set of file digital image sets;

(c) means for comparing one or more subframes of a queried digital image in the queried digital image set to one or more sets of one or more subframes of a file digital image in a file digital image set in the matching set of file digital image sets, wherein the comparing comprises comparing a set of ordered pixels in an RGB color space, whereby the set of ordered pixels and the RGB color space are common to the queried digital image and to the file digital image;

(d) means for translating, resizing, and requantizing the one or more subframes of the queried digital image and repeating (d) to reduce one or more bitmap matching errors of one or more subsets of the set of ordered pixels, wherein the one or more bitmap matching errors is taken between the queried digital image in a queried digital image set and the file digital image in a file digital image set in the matching set of file digital image sets;

(e) means for detecting a match if the one or more bitmap matching errors is less than a corresponding set of thresholds;

(f) means for adding, if no match is detected, to the one or more sets of one or more subframes of the file digital image, one of;

(1) one or more new subframes, and (2) a new set of one or more subframes; and

(g) means for repeating (c)-(f) for one or more file digital images in one or more file digital image sets in the matching set of file digital image sets.