Interpretation maps of deep neural networks are provided that use Renyi differential privacy to guarantee the robustness of the interpretation. In one aspect, a method for generating interpretation maps with guaranteed robustness includes: perturbing an original digital image by adding Gaussian noise to the original digital image to obtain m noisy images; providing the m noisy images as input to a deep neural network; interpreting output from the deep neural network to obtain m noisy interpretations corresponding to the m noisy images; thresholding the m noisy interpretations to obtain a top-k of the m noisy interpretations; and averaging the top-k of the m noisy interpretations to produce an interpretation map with certifiable robustness.

An image synthesis method includes inputting an original image and a watermark image into a synthesis model and obtaining a synthesized image output from the synthesis model. The original image and the watermark image are respectively processed in first and second sub-models of the synthesis model and then combined, and the concatenated result is processed in a third sub-model to generate the synthesized image.

A watermark representing a link to an original video and/or metadata such as haptic metadata associated with the original video is embedded in the original video in such a way that a re-recording to the original video can still preserve the watermark. The watermark can be used to link to the original video or to the metadata related thereto.

A method, computer system, and a computer program product for image protection are provided. A first device includes a computer and a first camera. The first device identifies a picture that has a hallmark. The first device transmits a message requesting authorization to have a copy of the picture. The message contains data representative of the hallmark. The first device receives a response message. The first device determines whether the response message authorizes capture of a copy of the picture.

Computer-implemented methods and systems are provided for digitally signing predetermined arrays of digital data. Such a method may provide a secret neural network model trained to classify arrays of digital data in dependence on data content of the arrays. The array of the arrays may be signed by supplying the array to the secret neural network model to obtain an initial classification result; and effecting a modification of data in the array to change the initial classification result to a predetermined, secret classification result, the modification being effected via a backpropagation process in the secret neural network model to progressively modify the array in response to backpropagated errors dependent on a difference between a current classification result for the array and the secret classification result.

The present disclosure relates to advanced image processing and encoded signal processing. One claim currently recites an image processing method comprising the acts: receiving a digital representation of artwork, the artwork having an area of uniform color; generating a two-dimensional data signal that redundantly encodes a plural-bit message, the data signal comprising plural elements, each of which has a single bit value; receiving a two-dimensional synchronization signal comprising plural elements, each of which has a plural-bit value, each element of said data signal having an element of the synchronization signal corresponding thereto; processing the two-dimensional data signal with the two-dimensional synchronization signal and with a gradient function to yield a two-dimensional gradient marking signal; and printing an ink counterpart of the gradient marking signal on a medium with the artwork, said printing comprising printing plural dithered two-dimensional blocks of at least four contiguous elements each, in which one or more elements of each block are printed to be dark. Of course, other claims and combinations are described as well.

This disclosure relates to advanced signal processing technology including signal encoding. One combination includes an apparatus comprising: memory for storing image data, the image data comprising a plurality of color separations or channels, in which the image data comprises at least a first type of machine-readable symbology comprising a 1D barcode represented therein and a second type of machine-readable symbology comprising a first signal represented therein, in which the second type of machine-readable symbology comprises a different type of machine-readable symbology relative to the first type of machine-readable symbology, the 1D barcode comprising a first plural-bit code and the first signal comprising a second plural-bit code; a barcode reader configured to analyze the image data to decode the 1D barcode to obtain the first plural-bit code; a signal decoder configured to analyze one or more color separations or channels of the plurality of color separations or channels to decode the first signal to obtain the second plural-bit code; one or more processors configured to determine whether the second plural-bit code and the first plural-bit code conflict; and to identify a conflict based on a conflict determination. Of course, other features and combinations are described as well.

This disclosure relates to advanced image signal processing technology including encoded signals and digital watermarking. We disclose methods, systems and apparatus for selecting which ink(s) should be selected to carry an encoded signal for a given machine-vision wavelength for a retail package or other printed design. We also disclose retail product packages and other printed objects, and methods to generate such, including a sparse mark in a first ink and an overprinted ink flood in a second ink. The first ink and the second ink are related through tack and spectral reflectance difference. Of course, other methods, packages, objects, systems and apparatus are described in this disclosure.

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.

A video player for playing a video stream that receives a master playlist identifying at least one variant playlist identifying a video file encoded as a series of video frames that when decoded provide the video stream. The video player based upon a configuration tag in the master playlist selectively determining whether the video file is to be processed in a trusted execution environment. The trusted execution environment of the video player selectively includes at least one of (i) hack one, only hack one; (ii) output and link protection; (iii) hardware root of trust; and (iv) forensic watermarking, and decrypts and/or decodes the video stream in such an environment.