A camera captures video imagery depicting a digitally-watermarked object. A reference signal in the watermark is used to discern the pose of the object relative to the camera, and this pose is used in affine-transforming and positioning a graphic on the imagery as an augmented reality overlay. Feature points are also discerned from the captured imagery, or recalled from a database indexed by the watermark. As the camera moves relative to the object, the augmented reality overlay tracks the changing object depiction, using these feature points. When feature point-based tracking fails, the watermark is again processed to determine pose, and the overlay presentation is updated accordingly. In another arrangement, feature points are extracted from images of supermarket objects captured by multiple users, and are compiled in a database in association with watermark data identifying the objects—serving as a crowd-sourced repository of feature point data. A great number of other features and arrangements are also detailed.

An example device for processing image data includes a memory configured to store an image; and one or more processors implemented in circuitry and configured to: process the image to identify a pilot signal in the image indicating a portion of the image, the pilot signal forming a boundary around the portion and having pixel values defined according to a mathematical relationship with pixel values within the portion such that the pilot signal is not perceptible to a human user and is detectable the device; determine the portion of the image using the pilot signal; and further process the portion to attempt to detect one or more contents of the portion without attempting to detect the one or more contents of the image in portions of the image outside the portion.

In one embodiment, a computing system may receive a video including a sequence of frames. The computing system may access a three-dimensional mask that specifies pixel-sampling locations, the three-dimensional mask having a first dimension and a second dimension corresponding to a spatial domain and a third dimension corresponding to a temporal domain. Blue noise property may be present in the pixel-sampling locations that are associated with each of a plurality of two-dimensional spatial slices of the three-dimensional mask in the spatial domain and the pixel-sampling locations that are associated with each of a plurality of one-dimensional temporal slices of the three-dimensional mask in the temporal domain. The computing system may generate a sample of the video by sampling the sequence of frames using the three-dimensional mask.

A system configured for implementing source identifiers into 3D content is configurable to access one or more numbers defining one or more visual characteristics of a 3D model and modify digits of the one or more numbers to include an identifier code. The identifier code is associated with a particular source such that subsequent detection of the identifier code within the 3D model or a derivative 3D model derived from the 3D model indicates that the 3D model or the derivative 3D model originated from the particular source.

A method is provided for watermarking a machine learning model used for object detection or image classification. In the method, a first subset of a labeled set of ML training samples is selected. The first subset is of a predetermined class of images. In one embodiment, the first pixel pattern is selected and sized to have substantially the same dimensions as each sample of the first subset or each bounding box in the case of an object detector. Each sample of the first subset is relabeled to have a different label than the original label. An opacity of the pixel pattern may be adjusted independently for different parts of the pattern. The ML model is trained with the labeled set of ML training samples and the first subset of relabeled ML training samples. Using multiple different opacity factors provides both reliability and credibility to the watermark.

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.

Optical code signal components are generated and then transformed into signal bearing art that conveys machine readable data. The components of an optical code are optimized to achieve improved signal robustness, reliability, capacity and/or visual quality. An optimization program can determine spatial density, dot distance, dot size and signal component priority to optimize robustness. An optical code generator transforms tiles of an optical code or image embedded with the optical code into signal-bearing art using stipple, Voronoi, Delaunay or other graphic drawing methods so as to retain prioritized components of the optical code. The optical code is merged into a host image, such as imagery, text and graphics of a package or label, or it may be printed by itself, e.g., on an otherwise blank label or carton. A great number of other features and arrangements are also detailed.

In one embodiment, a computing system may receive a video including a sequence of frames. The computing system may access a three-dimensional mask that specifies pixel-sampling locations, the three-dimensional mask having a first dimension and a second dimension corresponding to a spatial domain and a third dimension corresponding to a temporal domain. Blue noise property may be present in the pixel-sampling locations that are associated with each of a plurality of two-dimensional spatial slices of the three-dimensional mask in the spatial domain and the pixel-sampling locations that are associated with each of a plurality of one-dimensional temporal slices of the three-dimensional mask in the temporal domain. The computing system may generate a sample of the video by sampling the sequence of frames using the three-dimensional mask.

Some implementations provide a computer-assisted method for embedding information in an identification document. The method includes: modulating information by varying lengths of line segments; generating a line code that includes the line segments with corresponding lengths; and applying the line code on an identification document such that the information is used to authenticate the identification document as genuine when a person presents the identification document to verify an identify of the person.

An example device for processing image data includes a memory configured to store an image; and one or more processors implemented in circuitry and configured to: process the image to identify a pilot signal in the image indicating a portion of the image, the pilot signal forming a boundary around the portion and having pixel values defined according to a mathematical relationship with pixel values within the portion such that the pilot signal is not perceptible to a human user and is detectable the device; determine the portion of the image using the pilot signal; and further process the portion to attempt to detect one or more contents of the portion without attempting to detect the one or more contents of the image in portions of the image outside the portion.