US 7,609,849 B2Granted

Data processing apparatus and method

AssigneeSony United Kingdom Limited

InventorsDaniel Warren Tapson, Daniel Luke Hooper

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method includes generating, for a signal sample, first and second gaussian likelihood functions describing a likelihood of a watermark codeword coefficient value embedded into the signal sample being, respectively, positive and negative. The first gaussian likelihood function has a mean defined by a sum of a local mean value and a watermark strength value and a standard deviation defined by a local standard deviation. The second gaussian likelihood function has a mean defined by a difference between the local mean value and the watermark strength value and the standard deviation defined by the local standard deviation. The method also includes detecting a watermark in an image signal based on a probability of watermark codeword coefficients added to the signal sample being one of positive or negative, the probability being calculated from the first gaussian likelihood function and the second gaussian likelihood function.

Claims9

Dependent claims

A method of detecting a watermark codeword in a received image signal representing a watermarked image, using an image processing apparatus, the watermarked image comprising a plurality of signal samples, each of the plurality of signal samples representing a pixel value of the watermarked image, to which a watermark codeword coefficient value has been added, comprising:

low pass filtering the received image signal at the image processing apparatus to generate, for each of the plurality of signal samples, a local mean value;
subtracting the local mean value from the received image signal at the image processing apparatus to generate a residual signal comprising, for each of the plurality of signal samples, a residual value;
determining, from the residual signal, a local standard deviation for each of the plurality of signal samples at the image processing apparatus;
generating, with the image processing apparatus, for each of the plurality of signal samples, a watermark strength value providing an estimate of a magnitude with which the watermark codeword coefficient value has been added to the respective signal sample;
generating, with the image processing apparatus, for each of the plurality of signal samples, a first gaussian likelihood function describing a likelihood of the watermark codeword coefficient value embedded into the respective signal sample being positive, the first gaussian likelihood function having a mean defined by a sum of the local mean value and the watermark strength value and a standard deviation defined by the local standard deviation;
generating, with the image processing apparatus, for each of the plurality of signal samples, a second gaussian likelihood function describing a likelihood of the watermark codeword coefficient value embedded into the respective signal sample being negative, the second gaussian likelihood function having a mean defined by a difference between the local mean value and the watermark strength value and a standard deviation defined by the local standard deviation;
detecting with the image processing apparatus a watermark in the received image signal based on a probability of respective watermark codeword coefficients added to each of the plurality of signal samples being one of positive or negative, the probability being calculated from the first gaussian likelihood function and the second gaussian likelihood function.

4 more elements

A watermark detection apparatus for detecting a codeword in a received image signal representing a watermarked image, the watermarked image comprising a plurality of signal samples, each of the plurality of signal samples representing a pixel value of the watermarked image, to which a watermark codeword coefficient value has been added, the apparatus comprising:

a local mean generator operable to low pass filter the received image signal to generate, for each of the plurality of signal samples, a local mean value;
a residual generator operable to subtract the local mean value from the received image signal to generate a residual signal comprising, for each of the plurality of signal samples, a residual value;
a local standard deviation generator operable to determine from the residual signal, a local standard deviation for each of the plurality of signal samples;
a watermark strength estimator operable to generate for each of the plurality of signal samples, a watermark strength value providing an estimate of a magnitude with which the watermark codeword coefficient value has been added to the respective signal sample;
a likelihood generator operable to generate for each of the plurality of signal samples, a first gaussian likelihood function describing a likelihood of the watermark codeword coefficient value embedded into the respective signal sample being positive, the first gaussian likelihood function having a mean defined by a sum of the local mean value and the watermark strength value and a standard deviation defined by the local standard deviation, and to generate, for each of the plurality of signal samples, a second gaussian likelihood function describing a likelihood of the watermark codeword coefficient value embedded into the respective signal sample being negative, the second gaussian likelihood function having a mean defined by a difference between the local mean value and the watermark strength value and a standard deviation defined by the local standard deviation;
and
a watermark detector operable to detect a watermark in the received image signal based on a probability of respective watermark codeword coefficients added to each of the plurality of signal samples being one of positive or negative, the probability being calculated from the first gaussian likelihood function and the second gaussian likelihood function.

4 more elements

Dependencies are inferred from the claim text.

Facts

Application: 11/296,771Filed 7 Dec 2005
Priority: GB 0427033.69 Dec 2004
Published: US 2006/0153422 A113 Jul 2006
Granted: 27 Oct 2009
Term ends: Not estimatedThe source records don’t support an estimate.

Classification7 codes

G06T 1/005
Image watermarking: robust watermarking, e.g. average attack or collusion attack resistant
G06T 2201/0051
Image watermarking: embedding of the watermark in the spatial domain
G06T 2201/0052
Image watermarking: embedding of the watermark in the frequency domain
G06T 2201/0061
Image watermarking: embedding of the watermark in each block of the image, e.g. segmented watermarking

Family

US 2006/0153422 A1Pre-grant publication · published 13 Jul 2006
US 7,609,849 B2This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key