초록 ▼

A method and apparatus detects the presence of a watermark in digital data. The digital data may represent picture or sound information and maybe in the form of a broadcast television signal or a signal that has been recorded on a recording medium such as a compact disc. The watermark includes coeff

A method and apparatus detects the presence of a watermark in digital data. The digital data may represent picture or sound information and maybe in the form of a broadcast television signal or a signal that has been recorded on a recording medium such as a compact disc. The watermark includes coefficients which have been subject to an inverse local orthogonal transform before being embedded in the input data. In order to detect the presence of the watermark, the input watermarked data is first forward transformed and subtracted from the watermark coefficients so as to derive the data coefficients. The data coefficients are squared and formed into a local average to obtain a measure of the power in the local average. The watermark coefficients are divided by the local average so as to scale them and the scaled watermark coefficients are cross-correlated with the input data to detect whether the watermark is present. The cross-correlation is performed by means of a multiplier receiving the input data as a first input and the scaled watermark coefficients as a second input. The resulting output detection signal is subjected to a thresholding operation using a threshold set in dependence upon the global average power of the input data set.

대표청구항 ▼

1. A method of detecting the presence of a watermark in digital data, the watermark including coefficients embedded in the data, the method comprising the steps of, spectrally transforming the data and applying the transformed data as a first input and the watermark coefficients as a second input to

1. A method of detecting the presence of a watermark in digital data, the watermark including coefficients embedded in the data, the method comprising the steps of, spectrally transforming the data and applying the transformed data as a first input and the watermark coefficients as a second input to a cross correlator to provide a cross correlation, scaling the cross correlation by a local average characteristic of the data to form a detected correlation signal, forming a global average of said correlation signal, applying said global average of said correlation signal to one input of arithmetic means, forming a global average of the transformed data and applying the resultant to another input of said arithmetic means to form a detection signal. 2. A method as claimed in claim 1, wherein the step of scaling comprises scaling the watermark coefficients applied as the second input to the cross correlator. 3. A method as claimed in claim 1, wherein the step of scaling comprises scaling the transformed data applied as the first input to the cross correlator. 4. A method as claimed in claim 1, wherein the transformed data is applied to squaring means and averaged to form said local average characteristic. 5. A method as claimed in claim 1, wherein the scaling is performed by dividing the watermark coefficients by the local average characteristic. 6. A method as claimed in claim 1, wherein the detection signal is subjected to a thresholding operation. 7. A method as claimed in claim 1, wherein said arithmetic means comprises a square rooter and a divider. 8. Apparatus to detect the presence of a watermark in input digital data, the watermark including coefficients embedded in the input data, the apparatus comprising means to receive and spectrally transform the input data, a cross correlator having a first input to receive the transformed data and a second input to receive the watermark coefficients so as to form a cross-correlation signal, said cross-correlator including scaling means for scaling the transformed data by a local average characteristic of the data to form a detected correlation signal, means for forming a global average signal of said correlation signal and applying said global average signal to one input of arithmetic means, means for forming a global average of the transformed data and applying the resultant to another input of said arithmetic means whereby the output of the arithmetic means is a detection signal. 9. Apparatus as claimed in claim 8, wherein the scaling means is adapted to scale the cross correlation by scaling the watermark coefficients applied to the second input to the cross correlator. 10. Apparatus as claimed in8, wherein the scaling means is adapted to scale the cross correlation by scaling the input data applied to the first input to the cross correlator. 11. Apparatus as claimed in claim 8, further comprising a squarer to receive the transformed data which is averaged to form said local average characteristic. 12. Apparatus as claimed in claim 8, wherein the scaling means comprise a scaling divider to divide the watermark coefficients by the local average characteristic. 13. Apparatus as claimed in claim 8, wherein the cross correlator comprises a multiplier to multiply the scaled watermark coefficients with the transformed input data. 14. Apparatus as claimed in claim 8, further comprising thresholding means to subject the detection signal to a threshold. 15. Apparatus as claimed in claim 8, wherein the arithmetic means comprises a square rooter and a divider. 16. Apparatus to detect the presence of a watermark including coefficients embedded in digital input data, said apparatus including: means to receive and spectrally transform the input data to thereby form transformed data; local averaging means for providing a local average of the power of the transformed data at first and second outputs thereof; cross-correlation and scaling means connected to the first output of the local averag e means to scale the coefficients embedded in the input data and to cross-correlate the transformed data with the scaled coefficients to produce a correlation signal; first global averaging means to form a global average of the correlation signal; local average inverse means connected to the second output of the local average means so as to provide a local average inverse signal; second global averaging means to form a global average of the local average inverse means; and arithmetic means to receive the global average of the correlation signal and the local average inverse signal so as to form a detection signal. 17. Apparatus as claimed in claim 16 wherein the transformed data is applied to a squarer connected to an input of the local average means. 18. Apparatus as claimed in claim 16 wherein the cross-correlation and scaling means includes a divider to divide the coefficients by the local average of the power of the transformed data. 19. Apparatus as claimed in claim 18 wherein the cross-correlation and scaling means further include a multiplier to multiply the scaled coefficients with the transformed data. 20. Apparatus as claimed in claim 16 wherein the arithmetic means comprises a square rooter and a divider. 21. Apparatus as claimed in claim 16 further comprising thresholding means to subject the detection signal to a threshold. 22. A method of detecting the presence of a watermark in digital data, said watermark including coefficients embedded in the data, the method comprising the steps of: spectrally transforming the input data to form transformed data; providing a local average of the power of the transformed data at first and second outputs of a local average means; scaling the coefficients embedded in the data from the first output of the local average means and cross-correlating the transformed data with the scaled coefficients to thereby produce a correlation signal; providing a local average inverse signal of the second output of the local average means; forming a global average of the local average inverse signal; receiving the global average of the correlation signal and the local average inverse signal to form a detection signal.