System and method for detecting the watermark using decision fusion

• US 8,666,111 B2
• Filed: 07/28/2011
• Issued: 03/04/2014
• Est. Priority Date: 05/24/2011
• Status: Active Grant

• Current Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Original Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Inventors: Chalamala, Srinivasa Rao, Kakkirala, Krishna Rao, Dhillon, Jyoti, Venkataramani, Krithika
• Primary Examiner(s): Abdi, Amara
• Application Number:
  US13/193,122
• Publication Number:
  US 20120300975A1
• Time in Patent Office: 950 Days
• Field of Search: 382/100, 382/232, 380/51, 380/54, 380/287, 713/176
• US Class Current: 382/100
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06T 1/005 Robust watermarking, e.g. average attack or collusion attack resistant
• CPC Class Codes: G06T 2201/0065 Extraction of an embedded watermark; Reliable detection

Abstract

The present application provides a robust system and method for detecting the watermark in an electronic media, wherein the electronic media had gone through various kinds of attacks and their combinations thereof which may not be known while detecting the watermarks. The watermark detection system and method can be trained to detect or reject a particular pattern. The watermark detection system and method are based on correlation and are useful in any kind of digital watermarking applications. The watermark detection system and method perform well even when the amount of distortion is not precisely known, wherein the disclosure provides a set of templates or correlation filters, being designed for detection of watermark to cover any kind and combination of attacks. For synchronous attacks, the correlation filter designing is carried out dynamically. Particularly, the template i.e. correlation filter is an attack adaptive frequency domain pattern for watermark detection.

Claims

1. A method for detection of a plurality of attacks or distortions on an electronic media signal, the method comprising the computer implemented steps of:

extracting the electronic media signal from a compressed electronic media signal by employing at least one decompressing means, wherein the electronic media signal comprises one or more frames;

instantaneously estimating plurality of parameters associated with at least one distortion or attack on the electronic media signal;

estimating a type and a strength of distortions or attacks occurring simultaneously on the electronic media signal;

synchronously generating a set of candidate templates for each frame using the plurality of parameters, wherein each candidate template of the set of candidate templates is configured with a predictively designed cross correlation filter;

obtaining a Peak to Side lobe Ratio (PSR) score for each candidate template;

obtaining a maximum PSR score from the PSR score associated with each candidate template for each frame of the electronic media signal to obtain a classifier for each frame;

determining a correlation output between each candidate template and each frame of the electronic media signal based upon the PSR score;

detecting a watermark for each frame, using the classifier of each frame, when the maximum PSR score commensurates with a predefined threshold value; and

determining presence of the watermark in the electronic media signal using one or more fusion rules applied on the correlation output of each frame.

2. The method of claim 1, wherein the attacks or distortions comprise geometrical attacks, wherein the geometrical attacks further comprise a rotation, a scaling, a translation, a shearing, a random bending and change of aspect ratio, signal processing attacks comprising filtering and compression and cryptographic attacks comprising key search and oracle attacks.

3. The method of claim 1, wherein the electronic media signal comprises an audio, a video, an image, a speech, a computer readable medium, a Compact Disc (CD), a hard disk, a Floppy drive, and any kind of document or combination thereof.

4. The method of claim 1, wherein the estimation of the parameters is achieved by a feature based image matching algorithm adapted to estimate a plurality of parameters associated with the attacks or distortions, wherein the plurality of features comprises an amount of rotation, size scaling, cropping and a signal processing algorithm adapted to estimate a type and an amount of filtering used during the attacks or distortions.

5. The method of claim 1, wherein upon estimation of the plurality of parameters, the candidate templates are generated to configure possible combination of attacks or distortions until detection of a true positive rate.

6. The method of claim 1, wherein the decision fusion rule for detection of the watermark facilitates low false positive rates of the classifier associated with each candidate template.

15. The method of claim 1, further comprising applying at least one fusion rule for comparing a percentage of frames having the maximum PSR score associated with each frame for each candidate template to commensurate with a predefined threshold percentage.

16. The method of claim 1, further comprising applying at least one fusion rule for comparing an average of the maximum PSR scores associated with each frame for each candidate template to commensurate with the predefined threshold value thereof.

7. A system for detection of a plurality of attacks or distortions on an electronic media signal, the system comprising:

at least one decompressing means for extracting the electronic media signal from a compressed electronic media signal wherein the electronic media signal comprises one or more frames;

an estimation means adapted to instantaneously estimate a plurality of parameters associated with at least one distortion or attack on the electronic media signal;

a template generation means configured to

estimate a type and a strength of distortion or attacks occurring simultaneously on the electronic media signal; and

synchronously generate a set of candidate templates using the plurality of parameters, wherein each candidate template of the set of candidate templates is configured with a predictively designed cross correlation filter;

a correlation output determination means configured to

obtain a Peak to Side lobe Ratio (PSR) score for each candidate template;

obtain a maximum PSR score from the PSR score associated with each candidate template for each frame of the electronic media signal to obtain a classifier for each frame;

determine a correlation output between each candidate template and each frame of the electronic media signal based upon the PSR score; and

detect a watermark for each frame, using the classifier of each frame, when the maximum PSR score commensurates with a predefined threshold value; and

a watermark detection means configured to determine presence of the watermark in the electronic media signal using one or more fusion rules applied on the correlation output of each frame.

8. The system of claim 7, wherein the attacks or distortions comprise geometrical attacks, wherein the geometrical attacks further comprise rotation, a scaling, a translation, a shearing, a random bending, and a change of aspect ratio, signal processing attacks comprising filtering and compression and cryptographic attacks comprising key search and oracle attacks.

9. The system of claim 7, wherein the electronic media signal comprises an audio, a video, an image, a speech, a computer readable medium, a Compact Disc (CD), a hard disk, a Floppy drive, and any kind of document or combination thereof.

10. The system of claim 7, wherein the estimation means further comprises a feature based image matching algorithm adapted to estimate a plurality of parameters associated with the attacks or distortions, wherein the plurality of features comprises an amount of rotation, size scaling and cropping and a signal processing algorithm adapted to estimate a type and an amount of filtering used during the attacks or distortions.

11. The system of claim 7, wherein the correlation output determination means further comprises a score calculator adapted to instantaneously calculate a Peak to Side lobe Ratio (PSR) score associated with each candidate template for each frame of the electronic media signal.

12. The system of claim 7, wherein the decision fusion module provides a higher probability of detection of the watermark based on the correlation output.

13. The system of claim 7, wherein the predictively designed cross correlation filter comprises Random binary pattern trained to counter different attacks or distortions.

14. The system of claim 7, wherein the watermark detection means is adapted to detect the watermark for a plurality of encryption modality of the electronic media signals comprising encryption in a spatial domain, a frequency domain, and a time domain.

17. The system of claim 7, further comprising a decision fusion module configured to apply at least one fusion rule for comparing a percentage of frames having the maximum PSR score associated with each frame for each candidate template to commensurate with a predefined threshold percentage.

18. The system of claim 7, further comprising a decision fusion module configured to apply at least one fusion rule for comparing an average of the maximum PSR scores associated with each frame for each candidate template to commensurate with the predefined threshold value thereof.