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. A method comprising the acts:
applying data associated with input image data to a convolutional neural network;
receiving pose information output from the convolutional neural network; and
extracting a previously-encoded data payload from a pattern depicted in the input image data using the pose information to yield an extracted data payload.
2. The method of claim 1 in which said data associated with the input image data comprises pixel data, histogram data, L-tuple data, or oct-axis data.
3. The method of claim 2 that includes deriving histogram data from the data associated with the input image data, and applying said histogram data to the convolutional neural network.
4. The method of claim 2 that includes deriving L-tuple data from the data associated with the input image data, and applying said L-tuple data to the convolutional neural network.
5. The method of claim 2 that includes deriving oct-axis data from the data associated with the input image data, and applying said oct-axis data to the convolutional neural network.
6. The method of claim 1 in which the pose information indicates two more of rotation, scale or translation information.
7. The method of claim 1 that includes decoding a plural-bit payload from the extracted data payload.
8. The method of claim 1 that includes decoding a plural-bit payload from the extracted data payload.
9. The method of claim 8 that includes overlaying a visual augmentation on the input image data in accordance with the pose information and in accordance with the plural-bit payload.
10. The method of claim 1 that includes determining a distance to each of several patches on a surface of an object depicted in the input image data from the pose information.
11. The method of claim 10 that includes generating a 3D model of the object using said determined distances.
12. A non-transitory computer-readable medium having instructions that configure a programmable system to perform acts including:
applying data associated with input image data to a convolutional neural network;
receiving pose information output from the convolutional neural network; and
extracting a previously-encoded data payload from a pattern depicted in the input image data using the pose information to yield an extracted data payload.
13. A system including one or more hardware processors and associated memory, the one or more hardware processors being configured to:
apply data associated with input image data to a convolutional neural network;
receive pose information output from the convolutional neural network; and
extract a previously-encoded data payload from a pattern depicted in the input image data using the pose information to yield an extracted data payload.
14. The system of claim 13 in which the one or more hardware processors are further configured to implement said convolutional neural network, employing parameters and weights stored in said associated memory.
15. The system of claim 13 in which the data associated with the input image data comprises pixel data, histogram data, L-tuple data, or oct-axis data.
16. The system of claim 13 in which the one or more processors are configured to derive histogram data from the data associated with the input image data, and apply said histogram data to the convolutional neural network.
17. The system of claim 13 in which the one or more processors are configured to derive L-tuple data from the data associated with the input image data, and apply said L-tuple data to the convolutional neural network.
18. The system of claim 13 in which the one or more processors are configured to derive oct-axis data from the data associated with the input image data, and apply said oct-axis data to the convolutional neural network.
19. The system of claim 13 in which the pose information indicates two more of rotation, scale or translation information.
20. The system of claim 13 in which the one or more processors are configured to decode a plural-bit payload from the extracted data payload.
21. The system of claim 13 in which the one or more processors include an Application Specific Integrated Circuit (ASIC).
22. The system of claim 13 in which the one or more processors include a Field Programmable Gate Array (FPGA).