Image-based pose determination
A steganographic digital watermark signal is decoded from host imagery without requiring a domain transformation for signal synchronization, thereby speeding and simplifying the decoding operation. In time-limited applications, such as in supermarket point-of-sale scanners that attempt watermark decode operations on dozens of video frames every second, the speed improvement allows a greater percentage of each image frame to be analyzed for watermark data. In battery-powered mobile devices, avoidance of repeated domain transformations extends battery life. A great variety of other features and arrangements, including machine learning aspects, are also detailed.
1. An image processing method that includes the acts: receiving a two-dimensional (2D) array of image values, and applying a 2D filter kernel to the received 2D array of image values to produce a filtered array of image values as output data, characterized in that the 2D filter kernel comprises a center coefficient and plural non-center coefficients, wherein (a) the center coefficient of the 2D filter kernel has a first polarity and all of the plural non-center coefficients of the 2D filter kernel have a second, opposite polarity; (b) two of said coefficients are real-valued rather than integer-valued; and (c) the 2D filter kernel comprises three non-center coefficients having different values.
2. The method of claim 1 further characterized in that the 2D filter kernel comprises four non-center coefficients having different values.
3. The method of claim 1 further characterized in that the 2D filter kernel comprises five non-center coefficients having different values.
4. The method of claim 1 further characterized in that the 2D filter kernel comprises six non-center coefficients having different values.
5. The method of claim 1 further characterized in that the 2D filter kernel comprises seven non-center coefficients having different values.
6. The method of claim 1 further characterized in that the 2D filter kernel comprises eight non-center coefficients having different values.
7. The method of claim 1 that further includes applying a decoding process to the output data to extract information encoded in the 2D array of input values.
8. The method of claim 7 that includes extracting plural-symbol payload information encoded in the 2D array of input values as a digital watermark signal.
9. The method of claim 8 in which the 2D array of input values comprise pixels depicting both text and the digital watermark signal.
10. The method of claim 8 , performed by a point of sale scanning system in a retail store.
11. The method of claim 1 in which the absolute value of each of said coefficients is 0.923 or greater.
12. The method of claim 1 in which the absolute value of each of said coefficients is 1.125 or less.
13. The method of claim 1 in which the center coefficient has a value one-eighth that of the negative sum of the eight neighboring coefficients.
14. The method of claim 1 in which the plural non-center coefficients are symmetrical across a horizontal axis.
15. The method of claim 1 in which the plural non-center coefficients are symmetrical across vertical and horizontal axes.
16. The method of claim 1 in which the plural non-center coefficients are symmetrical across a diagonal axis.
17. The method of claim 1 in which four corner coefficients that neighbor the center coefficient have the same value.
18. An apparatus including a processor and associated memory, the processor being configured to receive a two-dimensional (2D) array of image values, and apply a 2D filter kernel to the received 2D array of image values to produce a filtered array of image values as output data, characterized in that the 2D filter kernel comprises a center coefficient and plural non-center coefficients, wherein (a) the center coefficient of the 2D filter kernel has a first polarity and all of the plural non-center coefficients of the 2D filter kernel have a second, opposite polarity; (b) two of said coefficients are real-valued rather than integer-valued; and (c) the 2D filter kernel comprises three non-center coefficients having different values.
19. A non-transitory computer readable medium containing software instructions that, when executed by a processor, cause the processor to receive a two-dimensional (2D) array of image values, and apply a 2D filter kernel to the received 2D array of image values to produce a filtered array of image values as output data, wherein the 2D filter kernel comprises a center coefficient and plural non-center coefficients, wherein (a) the center coefficient of the 2D filter kernel has a first polarity and all of the plural non-center coefficients of the 2D filter kernel have a second, opposite polarity; (b) two of said coefficients are real-valued rather than integer-valued; and (c) the 2D filter kernel comprises three non-center coefficients having different values.