Optimization methods for the insertion, protection, and detection of digital watermarks in digitized data

DC CAFC

US 5,889,868 A
Filed: 07/02/1996
Issued: 03/30/1999
Est. Priority Date: 07/02/1996
Status: Expired due to Term

- Alert
- Pin

Associated Cases

Associated Defendants

First Claim

Patent Images

1. A method for amplitude independent encoding of digital watermark information in a signal, comprising steps of:

determining in said signal a sample window having a minimum and a maximum;

determining a quantization interval of said sample window, where said quantitization interval can be used to quantize normalized window samples;

normalizing the sample window to provide normalized samples, where normalized samples conform to a limited range of values, proportional to real sample values, and comprise a representation of the real sample values with a resolution higher than the real range of values, and where the normalized values can be divided by the quantization interval into distinct quantization levels;

analyzing the normalized samples to determine quantization levels;

comparing the message bits to the corresponding quantization level information from the analyzing step;

when a bit conflicts with the quantization level, adjusting the quantization level of said sample window to correspond to the message bit; and

de-normalizing the analyzed normalized samples.

View all claims

6 Assignments

Timeline View

Assignment View

Litigations

0 Petitions

Accused Products

Abstract

The implementations of digital watermarks can be optimally suited to particular transmission, distribution and storage mediums given the nature of digitally-sampled audio, video and other multimedia works. Watermark application parameters can be adapted to the individual characteristics of a given digital sample stream. Watermark information can be either carried in individual samples or in relationships between multiple samples, such as in a waveform shape. More optimal models may be obtained to design watermark systems that are tamper-resistant given the number and breadth of existent digitized sample options with different frequency and time components. The highest quality of a given content signal may be maintained as it is mastered, with the watermark suitably hidden, taking into account usage of digital filters and error correction. The quality of the underlying content signals can be used to identify and highlight advantageous locations for the insertion of digital watermarks. The watermark is integrated as closely as possible to the content signal, at a maximum level to force degradation of the content signal when attempts are made to remove the watermarks.

810 Citations

57 Claims

1. A method for amplitude independent encoding of digital watermark information in a signal, comprising steps of:
- determining in said signal a sample window having a minimum and a maximum;
  
  determining a quantization interval of said sample window, where said quantitization interval can be used to quantize normalized window samples;
  
  normalizing the sample window to provide normalized samples, where normalized samples conform to a limited range of values, proportional to real sample values, and comprise a representation of the real sample values with a resolution higher than the real range of values, and where the normalized values can be divided by the quantization interval into distinct quantization levels;
  
  analyzing the normalized samples to determine quantization levels;
  
  comparing the message bits to the corresponding quantization level information from the analyzing step;
  
  when a bit conflicts with the quantization level, adjusting the quantization level of said sample window to correspond to the message bit; and
  
  de-normalizing the analyzed normalized samples.
- View Dependent Claims (2, 31, 32, 33, 34, 35, 36, 37, 38, 40, 41)
- - 2. The method according to claim 1, wherein watermark signal characteristics or a watermark certificate can be compressed.
  - 31. The method of claim 1, further comprising the step of randomly encoding watermark bits in the signal, wherein less than a total number of available candidate watermark bits are encoded as determined by at least one of a primary mask, a convolution mask, and a message delimiter mask.
  - 32. The method of claim 31, wherein at least one of the primary mask, the convolution mask, and the message delimiter mask comprise an optimal digital filter.
  - 33. The method of claim 32, further comprising the step of encoding additional bits in the signal using a separate predetermined key.
  - 34. The method of claim 31, further comprising the step of decorrelating additional bits using a dither technique.
  - 35. The method of claim 1, further comprising the step of randomly encoding message delimiter bits in the signal, wherein less than a total number of available candidate message delimiter bits are encoded as determined by at least one of a primary mask, a convolution mask, and a message delimiter mask.
  - 36. The method of claim 35, wherein at least one of the primary mask, the convolution mask, and the message delimiter mask comprise an optimal digital filter.
  - 37. The method of claim 36, further comprising the step of encoding additional bits in the signal using a separate predetermined key.
  - 38. The method of claim 35, further comprising the step of decorrelating additional bits using a dither technique.
  - 40. The method of claim 1, further comprising the step of adding one or more random noise bits to a set of bits representing a watermark before encoding the watermark.
  - 41. The method of claim 1, further comprising the step of randomly varying two or more adjacent sample windows.

3. A method for amplitude independent decoding of digital watermark information in a signal comprising steps of:
- determining in said signal a sample window having a minimum and a maximum;
  
  determining a quantization interval of said sample window, where said quantitization interval can be used to quantize normalized window samples;
  
  normalizing the sample window to provide samples, where normalized samples conform to a limited range of values, proportional to real sample values, and comprise a representation of the real sample values with a resolution higher than the real range of values, and where the normalized values can be divided by the quantization interval into distinct quantization levels; and
  
  analyzing the quantization level of said samples to determine a messagebit value.
- View Dependent Claims (4, 43, 44, 45)
- - 4. The method according to claim 3, wherein watermark signal characteristics or a watermark certificate can be compressed.
  - 43. The method of claim 3, further comprising the step of changing a transform coefficient amplitude.
  - 44. The method of claim 3, wherein a watermark bit occupies more time than a single transform frame.
  - 45. The method of claim 44, wherein a redundant block code is used to encode watermark bits, such that n bits are encoded into a block having a length of m bits, where m is several times greater than n.

5. A method of encoding and decoding watermarks in a signal, comprising insertion and detection of abstract signal features in said signal to carry watermark information, wherein said abstract signal features are mathematical functions of the input sample window, and by extension, adjacent sample windows.
- View Dependent Claims (42)
- - 42. The method of claim 5, further comprising the step of randomly varying two or more adjacent sample windows.

6. A method of pre-analyzing a digital signal for encoding digital watermarks using a digital filter comprising determining what changes in the digital signal will be affected by the digital filter.
- View Dependent Claims (7, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57)
- - 7. The method according to claim 6, further comprising a step of encoding watermarks so as to either avoid frequency or time delimited areas of the signal which will be changed by the digital filter, or ensure that the watermark will survive the changes instroduced by the digital filter.
  - 46. The method of claim 6, further comprising the step of encoding a watermark using a psychoacoustic model by first determining where watermark bits are inaudibly suited without introducing audible artifacts using a psychoacoustic filter, and then encoding the watermark bits by making signal modifications that are below a predicted masking level.
  - 47. The method of claim 46, wherein the watermark bits are encoded into a subset of time regions below the predicted masking level.
  - 48. The method of claim 46, wherein the watermark bits are encoded into a subset of frequency regions below the predicted masking level.
  - 49. The method of claim 7, further comprising the step of encoding a watermark using a psychoacoustic model by first determining where watermark bits are inaudibly suited without introducing audible artifacts using a psychoacoustic filter, and then encoding the watermark bits by making signal modifications that are below a predicted masking level.
  - 50. The method of claim 49, wherein the watermark bits are encoded into a subset of time regions below the predicted masking level.
  - 51. The method of claim 49, wherein the watermark bits are encoded into a subset of frequency regions below the predicted masking level.
  - 52. The method of claim 6, further comprising the step of encoding a watermark using a psychovisual model by first determining where watermark bits are invisibly suited without introducing visible artifacts using a psychovisual filter, and then encoding the watermark bits by making signal modifications that are below a predicted masking level.
  - 53. The method of claim 52, wherein the watermark bits are encoded into a subset of time regions below the predicted masking level.
  - 54. The method of claim 52, wherein the watermark bits are encoded into a subset of frequency regions below the predicted masking level.
  - 55. The method of claim 7, further comprising the step of encoding a watermark using a psychovisual model by first determining where watermark bits are invisibly suited without introducing visible artifacts using a psychovisual filter, and then encoding the watermark bits by making signal modifications that are below a predicted masking level.
  - 56. The method of claim 55, wherein the watermark bits are encoded into a subset of time regions below the predicted masking level.
  - 57. The method of claim 55, wherein the watermark bits are encoded into a subset of frequency regions below the predicted masking level.

8. A method of error coding watermark message certificates using cross interleaved codes which use error codes of high redundancy, including codes with Hamming distances of greater than or equal to n, wherein n is a number of bits in a message block.

9. A method of pre-processing a watermark message certificate comprising determining an exact length of the watermark message as it will be encoded.
- View Dependent Claims (10)
- - 10. The method according to claim 9, further comprising a step of generating a watermark key which will provide at least one unique bit for each bit comprising the watermark message.

11. A method of encoding a watermark in a digital signal, comprising the steps of:
- generating watermark pseudo-random key bits using a non-linear generator; and
  
  encoding the watermark in the digital signal using the watermark pseudo-random key bits and characteristics of the digital signal.

12. A method of encoding a watermark in a digital signal, comprising the steps of:
- generating watermark pseudo-random key bits using a chaotic generator; and
  
  encoding the watermark in the digital signal using the watermark pseudo-random key bits and characteristics of the digital signal.

13. A method of encoding a watermark in a digital signal, comprising the steps of:
- mapping pseudo-random key and processing state information to effect an encode/decode map using a non-linear generator; and
  
  encoding the watermark in the digital signal using the encode/decode map and characteristics of the digital signal.

14. A method of encoding a watermark in a digital signal, comprising the steps of:
- mapping pseudo-random key and processing state information to effect an encode/decode map using a chaotic generator; and
  
  encoding the watermark in the digital signal using the encode/decode map and characteristics of the digital signal.

15. A method of guaranteeing watermark certificate uniqueness comprising attaching a timestamp or user identification dependent hash or message digest of watermark certificate data to the certificate.
- View Dependent Claims (39)
- - 39. The method of claim 15, further comprising the step of using a predetermined key to decode message delimiter bits for verification of one or more digital signature bits.

16. A method of generating and modulating a local noise signal to contain watermark information, wherein the noise signal is a function of at least one variable which depends on key and processing state information.

17. A method of varying watermark quantizations such that the dither changes an absolute quantization value, but does not change a quantization level or information carried in the quantization.

18. A method of encoding watermarks comprising steps of:
- inverting at least one instance of the watermark bit stream; and
  
  p1 encoding at least one instance of the watermark using said inverted instance of the watermark bit stream.

19. A method of decoding watermarks comprising steps of:
- considering an original watermark synchronization marker, an inverted watermark synchronization marker, and inverted watermarks; and
  
  decoding based on the considering step.

20. A method of encoding and decoding watermarks in a signal using a spread spectrum technique to encode or decode where information is encoded or decoded at audible levels and the encoding and decoding methods are pseudo-random over frequency.

21. A method of encoding and decoding watermarks in a signal using a spread spectrum technique to encode or decode where information is encoded or decoded at audible levels and the encoding and decoding methods are pseudo-random over time.
- View Dependent Claims (22)
- - 22. The method of claim 21, wherein the information is encoded or decoded at audible levels and the encoding and decoding methods are pseudo-random, over both frequency and time.

23. A method of analyzing composite digitized signals for watermarks comprising steps of:
- obtaining a composite signal;
  
  obtaining an unwatermarked sample signal;
  
  time aligning the unwatermarked sample signal to the composite signal;
  
  gain adjusting the time aligned unwatermarked sample signal to a corresponding segment of the composite signal, determined in the time aligning step;
  
  estimating a pre-composite signal using the composite signal and the gain adjusted unwatermarked sample signal;
  
  estimating a watermarked sample signal by subtracting the estimated pre-composite signal from the composite signal; and
  
  scanning the estimated watermarked sample signal for watermarks.
- View Dependent Claims (28, 29, 30)
- - 28. The method of claim 23, further comprising the step of accessing spectral coefficient amplitude information in the watermarked sample signal.
  - 29. The method of claim 28, wherein the spectral coefficient amplitude information represents less than a total number of bits in the unwatermarked sample signal.
  - 30. The method of claim 28, wherein the spectral coefficient amplitude information represents a signal characteristic parameter for use in watermark decoding.

24. A method for varying watermark encode/decode algorithms automatically during the encoding or decoding of a watermark comprising steps of:
- a) assigning a list of desired CODECs to a list of corresponding signal characteristics which indicate use of particular CODECs;
  
  b) during encoding/decoding, analyzing characteristics of the current sample frame in the signal stream, prior to delivering the frame to a CODEC;
  
  c) looking up the corresponding CODEC from the list of CODECs in step (a) which matches the observed signal characteristics from step (b);
  
  d) loading and/or preparing the desired CODEC;
  
  e) passing the sample frame to the CODEC selected in step (c); and
  
  f) receiving the output samples from step (e).
- View Dependent Claims (25)
- - 25. The method according to claim 24, wherein watermark signal characteristics or a watermark certificate can be compressed.

26. A method for varying watermark encode/decode algorithms automatically during the encoding or decoding of a watermark comprising steps of:
- a) assigning a list of desired CODECs to a list of index values which correspond to values computed as a function of the pseudo-random watermark key and the state of the processing framework;
  
  b) during encoding/decoding, computing the pseudo-random key index value for the current sample frame in the signal stream, prior to delivering the frame to a CODEC;
  
  c) looking up the corresponding CODEC from the list of CODECs in step (a) which matches the index value from step (b);
  
  d) loading and/or preparing the desired CODEC;
  
  e) passing the sample frame to the CODEC selected in step (c); and
  
  f) receiving the output samples from step (e).
- View Dependent Claims (27)
- - 27. The method according to claim 26, wherein watermark signal characteristics or a watermark certificate can be compressed.

Specification

Resources

Litigation Campaign Assessment

Litigation Data

Current Assignee
Wistaria Trading Inc.
Original Assignee
The Dice Company
Inventors
Moskowitz, Scott A., Cooperman, Marc
Primary Examiner(s)
SWANN, TOD R

Application Number

US08/677,435
Time in Patent Office

1,001 Days
Field of Search

380/28, 380/46, 380/51, 380/4, 380/55
US Class Current

713/176
CPC Class Codes

H04L 2209/34   Encoding or coding, e.g. Hu...

H04L 2209/603   Digital right managament [DRM]

H04L 2209/608   Watermarking

H04L 9/3263   involving certificates, e.g...

H04N 21/23892   involving embedding informa...

H04N 21/8358   involving watermark protect...

Optimization methods for the insertion, protection, and detection of digital watermarks in digitized data

First Claim

6 Assignments

Litigations

0 Petitions

Accused Products

Abstract

810 Citations

57 Claims

Specification

Use Cases

Quick Links

Others

Optimization methods for the insertion, protection, and detection of digital watermarks in digitized data

First Claim

6 Assignments

Subscription Required

Subscription Required

Litigations

0 Petitions

Subscription Required

Accused Products

Subscription Required

Abstract

810 Citations

57 Claims

Specification

Subscription Required

Use Cases

Quick Links

Others