Method and apparatus for embedding data, including watermarks, in human perceptible sounds

• US 6,061,793 A
• Filed: 08/27/1997
• Issued: 05/09/2000
• Est. Priority Date: 08/30/1996
• Status: Expired due to Term

• Current Assignee: Digimarc Corporation
• Original Assignee: Regents of The University of Minnesota (University of Minnesota)
• Inventors: Boney, Laurence, Zhu, Bin, Tewfik, Ahmed H., Swanson, Mitchell D.
• Primary Examiner(s): Swann, Tod R.
• Assistant Examiner(s): Callahan, Paul E.
• Application Number:
  US08/918,891
• Time in Patent Office: 986 Days
• Field of Search: 360/18, 360/24, 360/27, 360/49, 348/461, 348/462, 381/124, 380/200, 380/206, 380/207, 380/237-238, 705/54, 704/216-218, 704/226-228, 704/500, 704/501, 704/503, 704/504, 369/4, 369/5, 386/95, 386/96, 386/104, 713/176
• US Class Current: 713/176
• CPC Class Codes: G10L 19/02 using spectral analysis, e.g. transform vocoders or subband vocoders
• CPC Class Codes: H04N 1/32165 using cosine transforms
• CPC Class Codes: H04N 1/3217 using wavelet transforms
• CPC Class Codes: H04N 1/32187 with selective or adaptive application of the additional information, e.g. in selected frequency coefficients
• CPC Class Codes: H04N 1/32192 according to calculated or estimated visibility of the additional information in the image
• CPC Class Codes: H04N 19/154 Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion use of rate-distortion criteria H04N19/147
• CPC Class Codes: H04N 21/23892 involving embedding information at multiplex stream level, e.g. embedding a watermark at packet level
• CPC Class Codes: H04N 21/8358 involving watermark protecting executable software by watermarking G06F21/16; image watermarking in general G06T1/0021; watermarks inserted in still images for transmission purposes H04N1/32144; inserting watermarks during video coding H04N19/467
• CPC Class Codes: H04N 2201/3233 of authentication information, e.g. digital signature, watermark
• CPC Class Codes: H04N 2201/327 which are undetectable to the naked eye, e.g. embedded codes

Abstract

A technique for hiding of data, including watermarks, in human-perceptible sounds, that is, audio host data, is disclosed. In one embodiment a method comprises three steps. In the first step, data to be embedded is inputted. In the case of a watermark, this data is a unique signature, and may be a pseudo-noise (PN) code. In the case of hidden data to be embedded in the host data, this data is the hidden data itself, or the hidden data as spread against the frequency spectrum by a pseudo-noise (PN) code. In the second step, the inputted data is embedded within the host data, in accordance with a perceptual mask of the host data. The perceptual mask determines the optimal locations within the host data to insert the inputted data. In the case of sounds, these optimal locations are determined by reference to the human auditory system. In the third step, the host data, with the embedded data, is further masked by a non-frequency mask. In the case of audio data, the non-frequency mask is a temporal mask.

Claims

1. A computerized method for embedding data into host data representing human-perceptible sounds comprising:

inputting the data;

embedding the data into the host data in accordance with a perceptual mask conducted in the frequency domain; and

,

further wherein the embedded data is embedded in accordance with a perceptual mask conducted in the temporal domain.

2. The computerized method of claim 1, wherein the data embedded into the host data comprises data representing a watermark acting as a unique identifier for the host data.

3. The computerized method of claim 2, wherein the watermark comprises a pseudo-noise (PN) sequence.

4. The computerized method of claim 1, wherein the data embedded into the host data comprises data spread against data representing a signature according to a spread-spectrum technique.

5. The computerized method of claim 4, wherein the signature comprises a pseudo-noise (PN) sequence.

18. The method according to claim 1 wherein the perceptual mask conducted in the temporal domain occurs prior to the perceptual mask conducted in the frequency domain.

19. The method according to claim 1 wherein the data to be embedded is adjusted according to the frequency masking thresholds and not the temporal masking thresholds.

20. The method according to claim 1 wherein the data to be embedded is adjusted according to the temporal masking thresholds and not the frequency masking thresholds.

21. The method according to claim 1 wherein the perceptual mask in the frequency domain is computed based on perceptual charactertistics of the host data frequency content.

22. The method according to claim 1 wherein the perceptual mask in the temporal domain is computed based on perceptual characteristics of the host data temporal content.

23. The method according to claim 1 wherein the embedding function consists of a nonlinear rounding operation which adjusts frequency and/or temporal samples of the host audio data by rounding the audio data according to a quantization threshold based on the corresponding frequency and/or temporal mask;

setting rounded value to a preset value to represent the embedded data bit; and

reconstructing the audio by reversing the quantization operation.

6. A computerized system for hiding hidden data having a plurality of bits within a host data representing human-perceptible sounds comprising:

a processor;

a computer-readable medium;

computer-executable instructions executed by the processor from the computer-readable medium comprising;

segmenting the host data into a plurality of blocks;

applying a discrete cosine transform (DCT) to each block of the host data to generate a frequency block corresponding to the block of host data;

applying a perceptual mask to each frequency block;

embedding each bit of the hidden data into a corresponding frequency block; and

,

applying an inverse DCT to each frequency block to generate the host data having the hidden data embedded therein.

7. The computerized system of claim 6, wherein the instructions further comprise spreading each bit of the hidden data by a signature after selecting a plurality of blocks of the host data.

8. The computerized system of claim 6, wherein the instructions further comprise applying a temporal mask to the host data having the hidden data embedded therein.

24. The method according to claim 6 where the discrete cosine transform (DCT), a specific frequency transform, is replaced by a general frequency transform.

9. A computer-readable medium having a computer program stored thereon to cause a suitably equipped computer to perform the method comprising:

segmenting host data representing human-perceptible sounds into a plurality of blocks;

spreading each of a plurality of bits of hidden data by a signature;

applying a discrete cosine transform (DCT) to each block of the host data to generate a frequency block corresponding to the block of host data;

applying a perceptual mask to each frequency block;

embedding each bit of the hidden data, as spread by the signature, into a corresponding frequency block;

applying an inverse DCT to each frequency block to generate the host data having the hidden data embedded therein; and

,

applying a temporal mask to the host data having the hidden data embedded therein.

10. The computer-readable medium of claim 9, wherein the medium is a floppy disk.

11. A computerized method for generating a watermark to embed into host data representing human-perceptible sounds comprising:

generating a watermark sequence;

calculating a masking threshold;

frequency-domain shaping the watermark sequence utilizing the masking threshold; and

,

time-domain weighting the watermark sequence after the watermark sequence has been frequency-domain shaped.

12. The computerized method of claim 11, wherein the watermark sequence comprises a PN-sequence.

13. The computerized method of claim 11, wherein the masking threshold is calculated using the MPEG Audio Psychoacoustic Model 1 or 2.

14. A computerized system for embedding a watermark into host data representing human-perceptible sounds comprising:

a processor;

a computer-readable medium;

a computer-executable instructions executed by the processor from the computer-readable medium comprising;

low-bit rate coding and decoding the host data;

generating a first interim watermark on the signal comprising of the host signal minus the low-bit rate coded and decoded host signal;

generating a second interim watermark comprising of the difference between the low-bit rate coded and decoded host signal minus the watermarked and low-bit rate coded and decoded host signal;

adding the first interim watermark to the second interim watermark to obtain the watermark; and

adding the watermark to the host data.

15. The computerized system of claim 14, wherein at least one of the first interim watermark and the second interim watermark is generated by performing a method comprising:

generating a watermark sequence;

calculating a masking threshold;

frequency-domain shaping the watermark sequence utilizing the masking threshold; and

,

time-domain weighting the watermark sequence after the watermark sequence has been frequency-domain shaped.

16. A computer-readable medium having a computer program stored thereon to cause a suitably equipped computer to perform the method comprising:

low-bit rate coding and decoding host data representing a human-perceptible sound;

generating a first interim watermark on the signal comprising of the host signal minus the low-bit rate coded and decoded host signal;

generating a second interim watermark comprising of the difference between the low-bit rate coded and decoded host signal minus the watermarked and low-bit rate coded and decoded host signal;

adding the first interim watermark to the second interim watermark to obtain the watermark; and

adding the watermark to the host data.

17. The computer-readable medium of claim 16, wherein the computer-readable medium is a floppy disk.

25. The method according to claim 16 wherein the perceptual mask conducted in the temporal domain occurs prior to the perceptual mask conducted in the frequency domain.

26. The method according to claim 16 wherein the data to be embedded is adjusted according to the frequency masking thresholds and not the temporal masking thresholds.

27. The method according to claim 16 wherein the data to be embedded is adjusted according to the temporal masking thresholds and not the frequency masking thresholds.

28. The method according to claim 16 wherein the perceptual mask in the frequency domain is computed based on perceptual charactertistics of the host data frequency content.

29. The method according to claim 16 wherein the perceptual mask in the temporal domain is computed based on perceptual characteristics of the host data temporal content.

30. The method according to claim 16 wherein the embedding function consists of a nonlinear rounding operation which adjusts frequency and/or temporal samples of the host audio data by rounding the audio data according to a quantization threshold based on the corresponding frequency and/or temporal mask;

setting rounded value to a preset value to represent the embedded data bit; and

reconstructing the audio by reversing the quantization operation.