Digital watermarking of acoustic data
Examples described herein include adding a digital watermark into acoustic data. For example, the digital watermark data can be embedded using a key such that the watermark data is spread across a spectrum of the acoustic data. Thus, the data for the digital watermark may have no or minimal visible impact in the images generated using the marked acoustic data. The digital watermark may provide improved security and privacy. Moreover, the digital watermark may be incorporated into the reconstruction process of images.
1. A method comprising:
providing acoustic data comprising time-series representation of acoustic echo signals from an object under inspection using an acoustic transducer array, the acoustic data being arranged in a 3D arrangement;
generating a first volume of slices based on the acoustic data, the first volume slices comprising 2D slices, wherein a parameter of the 2D slices is based on a number of elements of the acoustic transducer array;
for each of the first volume of slices:
transforming a respective slice into a frequency domain representation using a discrete transform to obtain a respective transformed slice,
embedding a portion of a digital watermark into the transformed slice, wherein the digital watermark is based on a key, the embedding comprising:
determining an energy level of the transformed slice, and
setting a strength coefficient for application of the key for the transformed slice based on the energy level of the transformed slice, and
applying an inverted form of the discrete transform to the transformed slice with the embedded portion of the digital watermark to generate a respective marked slice;
arranging respective marked slices with respective embedded portions of the digital watermark into a second volume of marked slices; and
generating an image based on the second volume of marked slices, wherein the image illustrates echo representations from the object.
2. The method of claim 1 , further comprises:
extracting the embedded digital watermark from the second volume of marked slices to generate a visual representation of the digital watermark.
3. The method of claim 2 , wherein extracting the embedded digital watermark includes
for each of a set of marked slices:
transforming a respective marked slice into the frequency domain representation to obtain a respective transformed marked slice,
determining a degree of correlation between the transformed marked slice and the key to generate a correlation result, and
based on the degree of correlation, establish a value of the portion of the recovered digital watermark;
aggregating the respective established values of portions of the recovered digital watermark; and
generating a visual representation of the recovered digital watermark based on the aggregated established values.
4. The method of claim 1 , wherein embedding the portion of the digital watermark into the transformed slice comprises:
based on a value of the portion of the digital watermark, combining the key with the transformed slice.
5. The method of claim 4 , further comprising:
based on the value of the portion of the digital watermark being a first type, adding the key to the transformed slice; and
based on the value of the portion of the digital watermark being a second type, subtracting the key from the transformed slice.
6. The method of claim 1 , wherein the key includes a bitmap representation.
7. The method of claim 6 , wherein the bitmap representation comprises at least one of a random pattern or random pixel values.
8. The method of claim 1 , wherein the digital watermark is a binary bitmap.
9. A method comprising:
providing acoustic echo data comprising time-series representations of acoustic echo signals from an object under inspection obtained using an acoustic transducer array;
aggregating the received acoustic data into a first three-dimensional matrix of echo data samples;
generating a set of two-dimensional slices from the matrix, wherein a parameter of the two-dimensional slices is based on a number of elements of the acoustic transducer array;
for each of the set of two-dimensional slices from the first matrix:
transforming a respective slice into a frequency domain representation using a discrete transform to obtain a respective transformed slice,
embedding a portion of a digital watermark into the transformed slice, the digital watermark based on a key, the embedding comprising:
determining an energy level of the transformed slice, and
setting a strength coefficient for application of the key for the transformed slice based on the energy level of the transformed slice, and
applying an inverted form of the discrete transform to the transformed slice with the embedded portion of the digital watermark to generate a respective marked slice;
arranging respective marked slices with respective embedded portions of the digital watermark into a second three-dimensional matrix of echo data samples; and
generating an image based on the second matrix, wherein the image illustrates echo representations from the object.
10. The method of claim 9 , wherein the matrix of echo data sample comprises a matrix of A-scan time-series representations of received acoustic echo data.
11. The method of claim 9 , wherein the generating the image comprises applying a total focusing method (TFM) imaging technique to the second volume of marked slices to generate the image.
12. The method of claim 9 , further comprises
extracting the embedded portions of the digital watermark from the second matrix to generate a visual representation of the digital watermark.
13. A system comprising:
one or more processors of a machine; and
a memory storing instructions that, when executed by the one or more processors, cause the machine to perform operations comprising:
providing acoustic data comprising time-series representation of acoustic echo signals from an object under inspection using an acoustic transducer array, the acoustic data being arranged in a 3D arrangement;
generating a first volume of slices based on the acoustic data, the first volume slices comprising 2D slices, wherein a parameter of the 2D slices is based on a number of elements of the acoustic transducer array;
for each of the first volume of slices:
transforming a respective slice into a frequency domain representation using a discrete transform to obtain a respective transformed slice,
embedding a portion of a digital watermark into the transformed slice, wherein the digital watermark is based on a key, the embedding comprising:
determining an energy level of the transformed slice, and
setting a strength coefficient for application of the key for the transformed slice based on the energy level of the transformed slice, and
applying an inverted form of the discrete transform to the transformed slice with the embedded portion of the digital watermark to generate a respective marked slice;
arranging respective marked slices with respective embedded portions of the digital watermark into a second volume of marked slices; and
generating an image based on the second volume of marked slices, wherein the image illustrates echo representations from the object.
14. The system of claim 13 , the operations further comprises:
extracting the embedded digital watermark from the second volume of marked slices to generate a visual representation of the digital watermark.
15. The system of claim 14 , wherein extracting the embedded digital watermark includes
for each of a set of marked slices:
transforming a respective marked slice into the frequency domain representation to obtain a respective transformed marked slice,
determining a degree of correlation between the transformed marked slice and the key to generate a correlation result, and
based on the degree of correlation, establish a value of the portion of the recovered digital watermark;
aggregating the respective established values of portions of the recovered digital watermark; and
generating a visual representation of the recovered digital watermark based on the aggregated established values.
16. The system of claim 13 , wherein embedding the portion of the digital watermark into the transformed slice comprises:
based on a value of the portion of the digital watermark, combining the key with the transformed slice.
17. The system of claim 16 , the operations further comprising:
based on the value of the portion of the digital watermark being a first type, adding the key to the transformed slice; and
based on the value of the portion of the digital watermark being a second type, subtracting the key from the transformed slice.
18. The system of claim 13 , wherein the key includes a bitmap representation.
19. The system of claim 18 , wherein the bitmap representation comprises at least one of a random pattern or random pixel values.
20. The system of claim 13 , wherein the digital watermark is a binary bitmap.