Methods, apparatus and articles of manufacture (e.g., computer readable media) to detect watermark modifications are disclosed. Example apparatus disclosed herein include means for encoding a first watermark in a first media signal obtained from an output of a media device to obtain a second media signal encoded with the first watermark and a second watermark, the second watermark already encoded in the first media signal obtained from the output of the media device. Disclosed example apparatus also include means for decoding the first watermark and the second watermark from the second media signal to determine a first metric corresponding to the first watermark and a second metric corresponding to the second watermark. Disclosed example apparatus further include means for outputting, based on the first metric and the second metric, an indication of whether the second watermark has been modified.

This disclosure relates to counterfeit detection and deterrence using advanced signal processing technology including steganographic embedding and digital watermarking. Digital watermark can be used on consumer products, labels, logos, hang tags, stickers and other objects to provide counterfeit detection mechanisms.

Systems and methods are directed to a computing system. The computing system can include one or more processors, a message embedding model, a message extraction model, and a first set of instructions that cause the computing system to perform operations including obtaining the three-dimensional image data and the message vector. The operations can include inputting three-dimensional image data and a message vector into the message embedding model to obtain encoded three-dimensional image data. The operations can include using the message extraction model to extract an embedded message from the encoded three-dimensional image data to obtain a reconstructed message vector. The operations can include evaluating a loss function for a difference between the reconstructed message vector and the message vector and modifying values for parameters of at least the message embedding model based on the loss function.

Exemplary systems and methods are disclosed for detecting embedded data in a digital image. The system includes a processing device that extracts one or more features from a digital image and analyzes the one or more extracted features in a plurality of steganography analyzers, each steganography analyzer configured to execute a different steganography algorithm. The processing device generates an output data value at each steganography analyzer, the output data value indicating a probability that the digital image includes steganography according to the steganography algorithm of the steganography analyzer. Each output probability value is fed to an ensemble classifier, the ensemble classifier including a neural network in which the output probability values of the plurality of steganography analyzers are ensembled together to generate an output ensemble data value indicating a probability that the digital image includes any steganography according to the steganography algorithms of the steganography analyzers.

A digital watermark analysis apparatus, comprising an image capturing unit for capturing a printed product where additional information is embedded by superimposing a pattern on an image to acquire a captured image, a calculation unit for calculating a spatial frequency characteristic of each small area in the captured image, a specifying unit for specifying an embedded signal strength and an embedding position of the additional information, a decision unit for deciding, based on the embedding position information and the embedded signal strength information, a position in the captured image of a marker detection area for detecting a marker as a reference position for acquiring the additional information, and an acquisition unit for detecting a marker in the marker detection area and acquiring the additional information in the captured image with reference to the detected marker.

Digital watermarking is adapted for the variable data printing. A reference signal serves as a proxy for optimizing the embedding a watermark in a host image to be printed. Using the reference signal, embedding parameters are generated, which are a function of constraints such as visual quality and robustness of the machine readable data. Adjustments needed to embed a unique payload in each printed piece are generated using the embedding parameters. These adjustments are stored in a manner that enables them to be efficiently obtained and applied within the RIP or press during operation of the press. Various other methods, system configurations and applications are also detailed.

The present disclosure relates generally to data hiding for retail product packaging and other printed objects such as substrates. One embodiment embeds an information signal in a spot color for printing on various substrates. The spot color is screened, and overprinted with process color tint. The tint is modulated prior to overprinting with optimized signal tweaks. The optimization can include consideration of a detector spectral dependency (e.g., red and/or green illumination). Many other embodiments and combinations are described in the subject patent document.

Methods, systems, and computer programs encoded on a computer storage medium, that relate to extracting digital watermarks from images, irrespective of distortions introduced into these images. Methods can include inputting a first data item into a channel encoder that can generate a first encoded data item that is greater in length than the first data item and that (1) includes the input data item and (2) new data this is redundant of the input data item. Based on the first encoded data item and a first image, an encoder model can generate a first encoded image into which the first encoded data is embedded as a digital watermark. A decoder model can decode the first encoded data item to generate a second data, which can be decoded by the channel decoder to generate data that is predicted to be the first data.

Artwork carrying machine readable data is generated by editing artwork according to a data signal or transforming the data signal into artwork. The machine-readable data signal is generated from a digital payload and converted into an image tile. Artwork is edited according to the image tile by moving graphic elements, adapting intersections of lines, or altering line density, among other techniques. Artwork is generated from the data signal by skeletonizing it and applying morphological operators to a skeletal representation, such as a medial axis transform. Artistic effects are introduced by filtering the data signal with directional blurring or shape filters.

The present disclosure relates generally to signal encoding for elements within PDF files. One implementation encodes an artwork element under different encoding conditions, and selects a winner version based on resulting signal robustness and/or visibility. Other implementations generate PDF layer masks to help determine overall embedding robustness, including interference from layered elements. Other implementations are provided too.