Techniques and systems for embedding and detecting watermarks in digital data

• US 7,496,209 B2
• Filed: 07/27/2004
• Issued: 02/24/2009
• Est. Priority Date: 07/29/2003
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Zhang, Jian, Lu, Wei
• Primary Examiner(s): Bali; Vikkram
• Assistant Examiner(s): Park; Soo Jin
• Application Number:
  US10/900,738
• Publication Number:
  US 20050025337A1
• Time in Patent Office: 1,673 Days
• Field of Search: 382/100, 713/716, 380/54, 380/55
• US Class Current: 382/100
• CPC Class Codes: G06T 1/0028 Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
• CPC Class Codes: G06T 2201/0051 Embedding of the watermark in the spatial domain
• CPC Class Codes: G06T 2201/0052 Embedding of the watermark in the frequency domain
• CPC Class Codes: G06T 2201/0061 Embedding of the watermark in each block of the image, e.g. segmented watermarking
• CPC Class Codes: G06T 2201/0081 whereby both original and watermarked images are required at decoder, e.g. destination-based, non-blind, non-oblivious
• CPC Class Codes: G06T 2201/0083 whereby only watermarked image required at decoder, e.g. source-based, blind, oblivious
• CPC Class Codes: H04N 19/467 characterised by the embedded information being invisible, e.g. watermarking

Abstract

A method for embedding and detecting watermarks in digital data. Data is analysed by an independent component analysis to derive a transform matrix W encoding properties of the data. The data is encoded using this transform matrix W, and a watermark is embedded into it. Then the inverse of the transform matrix is applied to obtain watermarked data. The presence of the watermark is found by applying the transform matrix again, and examining the result for the presence of the watermark.

Claims

1. A method for embedding a watermark in data, including a processor to perform the steps of:

analysing the data to derive a transform matrix for extracting from the data amplitudes of statistically mutually independent components of the data;

encoding the data using the transform matrix;

classifying said components in accordance with their energies, into higher, lower and medium energy ranges;

selecting the components classified in said medium energy range;

modifying at least a portion of the selected components using a watermark pattern; and

decoding the data using an inverse of the transform matrix to obtain watermarked data in which the watermark pattern is embedded.

2. The method according to claim 1, wherein the components are obtained by an independent component analysis (ICA).

3. The method according to claim 2, wherein the ICA is performed using multiple patches of the data.

4. The method according to claim 3 including performing a principal component analysis of a matrix formed using the patches of the data.

5. The method according to claim 1, wherein the selected components are further selected based on a component similarity criterion.

6. The method according to claim 1, wherein the data is image data, and the watermarked data is watermarked image data.

7. The method according to claim 6, wherein the image data is obtained by decompressing compressed image data, the method further including a step of compressing the watermarked image data.

8. The method according to claim 1, wherein the data is compressed image data.

9. The method according to claim 1 further comprising a step of generating the watermark pattern using watermark data and a key for generating at least one pseudorandom pattern, the pseudorandom pattern being modulated with the watermark data.

10. The method according to claim 1, wherein the selected components are modified by adding the watermark pattern linearly to it.

11. The method according to claim 1, wherein the selected components are modified to make the values of groups of two or more items of the encoded data differ by values which are in predetermined ranges.

12. The method of detecting a presence of the watermark pattern in watermarked data derived from a method according to claim 1, the detecting method comprising:

an encoding step of encoding the watermarked data using a transform matrix encoding properties of the watermarked data; and

an examination step of using a watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern.

13. The method according to claim 12 further comprising a step before the examination step of generating the watermark pattern from watermark data and a key.

14. The method according to claim 12, wherein the part of the encoded watermarked data is selected based on properties of the encoded watermarked data.

15. The method according to claim 12, wherein the examination step includes calculating a correlation between the watermark data and the part of the encoded watermarked data, and determining if the correlation is above a threshold.

16. The method according to claim 12, wherein the examination step includes comparing the values of different ones of the predetermined items of the encoded watermarked data.

17. The method according to claim 12, wherein the examination step includes subtracting from the encoded watermarked data corresponding values of encoded data obtained in the encoding step of claim 1.

18. An apparatus for detecting a presence of the watermark pattern in watermarked data derived from a method according to claim 1, the apparatus comprising a processor arranged to:

encode the watermarked data using a transform matrix encoding properties of the watermarked data; and

use a watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern.

19. The apparatus according to claim 18, wherein the processor is arranged, before using the watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern, to generate the watermark pattern from watermark data and a key.

20. The apparatus according to claim 18, wherein the part of the encoded watermarked data is selected based on properties of the encoded watermarked data.

21. The apparatus according to claim 18, wherein the processor is arranged to use the watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern by calculating a correlation between the watermark data and the part of the encoded watermarked data, and determining if the correlation is above a threshold.

22. The apparatus according to claim 18, wherein the processor is arranged to use the watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern by comparing the values of different ones of the predetermined items of the encoded watermarked data.

23. The apparatus according to claim 18, wherein the processor is arranged to subtract from the encoded watermarked data corresponding values of encoded data obtained in the encoding step of claim 1 when using the watermark pattern to examine at least a part of the encoded watermarked data for the presence of the watermark pattern.

24. The apparatus according to claim 18, wherein the part of the encoded watermarked data is selected based on properties of the encoded watermarked data.

25. An apparatus for embedding a watermark in data, including a processor arranged to:

analyse the data to derive a transform matrix for extracting from the data amplitudes of components of the data which are statistically mutually independent;

encode the data using the transform matrix;

classify said components in accordance with their energies, into higher, lower and medium energy ranges;

select the components classified in said medium energy range;

modify at least a portion of the selected components using a watermark pattern; and

decode the data using an inverse of the transform matrix to obtain watermarked data in which the watermark pattern is embedded.

26. The apparatus according to claim 25, wherein the processor is arranged to obtain the components by an independent component analysis (ICA).

27. The apparatus according to claim 26, wherein the processor is arranged to perform the ICA using multiple patches of the data.

28. The apparatus according to claim 27, wherein the processor is arranged to perform a principal component analysis of a matrix formed using the patches of the data.

29. The apparatus according to claim 25, wherein the processor is arranged to further select the components based on a component similarity criterion.

30. The apparatus according to claim 25 for watermarking image data.

31. The apparatus according to claim 30 which includes decompression means for obtaining the image data by decompressing compressed image data, and compression means for compressing the watermarked image data.

32. The apparatus according to claim 25 further comprising a database for storing watermark data and a database holding a key, the processor being arranged to generate at least one pseudorandom pattern, and modulate the pseudorandom pattern using the watermark data.

33. The apparatus according to claim 25, wherein the selected components are modified by adding the watermark pattern linearly to it.

34. The apparatus according to claim 25, wherein the selected components are modified to make the values of groups of two or more items of the encoded data differ by values which are in predetermined ranges.

35. A method for embedding a watermark in data, including a processor to perform the steps of:

analysing the data to derive a transform matrix for extracting from the data amplitudes of statistically mutually independent components of the data;

encoding the data using the transform matrix;

modifying at least a portion of the encoded data with a watermark pattern; and

decoding the data using an inverse of the transform matrix to obtain watermarked data in which the watermark pattern is embedded,

wherein the modified portion of the encoded data is selected by ordering the components by their amplitudes and selecting components having amplitudes within one or more amplitude ranges, and is further selected by ordering the components according to their main properties, and selecting components having similar parameters of the main properties,

wherein the component main properties include spatial location, orientation and frequency magnitude.

36. An apparatus for embedding a watermark in data, including a processor arranged to:

analyse the data to derive a transform matrix for extracting from the data amplitudes of components of the data which are statistically mutually independent;

encode the data using the transform matrix;

modify at least a portion of the encoded data using a watermark pattern; and

decode the data using an inverse of the transform matrix to obtain watermarked data in which the watermark pattem is embedded,

wherein the modified portion of the encoded data is selected by ordering the components by their amplitudes and selecting components having amplitudes within one or more amplitude ranges, and is further selected by ordering the components according to their main properties, and selecting components having similar parameters of the main properties,

wherein the component main properties include spatial location, orientation and frequency magnitude.