Differential modulation for robust signaling and synchronization

• US 10,181,170 B2
• Filed: 08/25/2017
• Issued: 01/15/2019
• Est. Priority Date: 01/22/2015
• Status: Active Grant

• Current Assignee: Digimarc Corporation
• Original Assignee: Digimarc Corporation
• Inventors: Stach, John, Kamath, Ajith M.
• Primary Examiner(s): Fitzpatrick, Atiba O
• Application Number:
  US15/687,332
• Publication Number:
  US 20180047125A1
• Time in Patent Office: 508 Days
• Field of Search: None
• CPC Class Codes: G06T 1/0028 Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
• CPC Class Codes: G06T 1/005 Robust watermarking, e.g. average attack or collusion attack resistant
• CPC Class Codes: G06T 1/0064 Geometric transfor invariant watermarking, e.g. affine transform invariant
• CPC Class Codes: G06T 2201/0061 Embedding of the watermark in each block of the image, e.g. segmented watermarking
• CPC Class Codes: G06T 2201/0065 Extraction of an embedded watermark; Reliable detection
• CPC Class Codes: G06T 2201/0202 whereby the quality of watermarked images is measured; Measuring quality or performance of watermarking methods; Balancing between quality and robustness

Abstract

Differential modulation schemes encode a data channel within host signal or noisy environment in a manner that is robust, flexible to achieve perceptual quality constraints, and provides improved data capacity. Differential arrangements enable a decoder to suppress host signal or other background signal interference when detecting, synchronizing and extracting an encoded data channel. They also enable the incorporation of implicit or explicit synchronization components, which are either formed from the data signal or are complementary to it.

Claims

1. A method of decoding an auxiliary data signal embedded in a host image, the method comprising:

for each of plural filtering locations within the host image, applying an extraction filter that extracts auxiliary data signal elements from differential relationships between host image signal elements at non-adjacent embedding locations within a block of neighboring embedding locations around a filtering location, wherein the differential relationships are formed by differential modulation of the auxiliary data signal within the host image;

applying geometrically distorted versions of the extraction filter to the host image to obtain filter output;

obtaining geometric distortion parameters from correlation computed from the filter output;

applying the extraction filter at an orientation of the geometric distortion parameters to extract estimates of variable data elements of the auxiliary data signal; and

performing error correction decoding of the estimates of the variable data elements.

8. The method of claim 1 wherein the extraction filter comprises plural orthogonal filter patterns, each of the plural orthogonal filter patterns corresponding to different data channels.

2. The method of 1 wherein the applying of the geometrically distorted versions comprises:

convolving the geometrically distorted versions of the extraction filter with the host image.

3. The method of claim 2 wherein the correlation is computed from the filter output by summing correlations obtained by applying the extraction filter at the plural filtering locations, and obtaining the geometric distortion parameters comprises selecting a candidate orientation associated with a first extraction filter at a candidate orientation that provides a strongest correlation measure based on summing of correlations obtained by applying the first extraction filter at the candidate orientation.

4. The method of 2 wherein the applying of the geometrically distorted versions comprises convolving the geometrically distorted versions of the extraction filter with a pre-filtered version of the host image.

5. The method of 1 wherein the geometrically distorted versions of the extraction filter comprise the extraction filter geometrically distorted according to plural different pairs of rotation and scale parameters.

6. The method of 1 wherein the filter output comprises a feature vector whose elements each correspond to a signal metric computed by applying a geometrically distorted version of an extraction filter to the host image;

and obtaining the geometric distortion parameters comprises correlating the feature vector with pre-computed feature vectors, each pre-computed feature vector having corresponding geometric distortion parameters, and obtaining geometric distortion parameters as the geometric distortion parameters corresponding to a pre-computed feature vector that has a maximum correlation with the feature vector.

7. The method of 5 wherein the pre-computed feature vectors are derived by applying the geometrically distorted extraction filters to an image signal modulated with a message payload using a differential modulation pattern.

9. A non-transitory computer readable medium on which stored instructions, which when executed by a processor, perform a method of decoding an auxiliary data signal embedded in a host image, the method comprising:

for each of plural filtering locations within the host image, applying an extraction filter that extracts auxiliary data signal elements from differential relationships between host image signal elements at non-adjacent embedding locations within a block of neighboring embedding locations around a filtering location, wherein the differential relationships are formed by differential modulation of the auxiliary data signal within the host image;

applying geometrically distorted versions of the extraction filter to the host image to obtain filter output;

obtaining geometric distortion parameters from correlation computed from the filter output;

applying the extraction filter at an orientation of the geometric distortion parameters to extract estimates of variable data elements of the auxiliary data signal; and

performing error correction decoding of the estimates of the variable data elements.

10. The non-transitory computer readable medium of claim 9 wherein the applying of the geometrically distorted versions comprises:

convolving the geometrically distorted versions of the extraction filter with the host image.

11. The non-transitory computer readable medium of claim 10 wherein the applying of the geometrically distorted versions comprises convolving the geometrically distorted versions of the extraction filter with a pre-filtered version of the host image.

12. The non-transitory computer readable medium of claim 9 wherein the geometrically distorted versions of the extraction filter comprise the extraction filter geometrically distorted according to plural different pairs of rotation and scale parameters.

13. The non-transitory computer readable medium of claim 9 wherein the filter output comprises a feature vector whose elements each correspond to a signal metric computed by applying a geometrically distorted version of an extraction filter to the host image;

and obtaining the geometric distortion parameters comprises correlating the feature vector with pre-computed feature vectors, each pre-computed feature vector having corresponding geometric distortion parameters, and obtaining geometric distortion parameters as the geometric distortion parameters corresponding to a pre-computed feature vector that has a maximum correlation with the feature vector.

14. The non-transitory computer readable medium of claim 13 wherein the pre-computed feature vectors are derived by applying the geometrically distorted extraction filters to an image signal modulated with a message payload using a differential modulation pattern.

15. An auxiliary data signal decoder comprising:

an image sensor;

memory configured to store images captured by the image sensor;

a processor in communication with the memory, the processor configured with instructions to;

apply an extraction filter that extracts auxiliary data signal elements from differential relationships between image signal elements at non-adjacent embedding locations within a block of neighboring embedding locations around a filtering location for each of plural filtering locations within a host image in the memory, wherein the differential relationships are formed by differential modulation of the auxiliary data signal within the host image;

apply geometrically distorted versions of the extraction filter to the host image to obtain filter output;

obtain geometric distortion parameters from correlation computed from the filter output;

apply the extraction filter at an orientation of the geometric distortion parameters to extract estimates of variable data elements of the auxiliary data signal; and

perform error correction decoding of the estimates of the variable data elements.

16. The auxiliary signal data decoder of claim 15 wherein the processor is configured to convolve the geometrically distorted versions of the extraction filter with the host image.

17. The auxiliary signal data decoder of claim 16 wherein the processor is configured to convolve the host image with a pre-filter.

18. The auxiliary signal data decoder of claim 15 wherein the geometrically distorted versions of the extraction filter comprise the extraction filter geometrically distorted according to plural different pairs of rotation and scale parameters.

19. The auxiliary signal data decoder of claim 15 wherein the filter output comprises a feature vector whose elements each correspond to a signal metric computed by applying a geometrically distorted version of an extraction filter to the host image;

and the processor is configured to correlate the feature vector with pre-computed feature vectors, each pre-computed feature vector having corresponding geometric distortion parameters, and obtain geometric distortion parameters as the geometric distortion parameters corresponding to a pre-computed feature vector that has a maximum correlation with the feature vector.

20. The auxiliary signal data decoder of claim 19 wherein the pre-computed feature vectors are derived by applying the geometrically distorted extraction filters to an image signal modulated with a message payload using a differential modulation pattern.