Systems, methods, and computer program products for determining an attack on a neural network. A data sample is received at a first classifier neural network and at a watermark classifier neural network, wherein the first classifier neural network is trained using a first dataset and a watermark dataset. The first classifier neural network determines a classification label for the data sample. A watermark classifier neural network determines a watermark classification label for the data sample. A data sample is determined as an adversarial data sample based on the classification label for the data sample and the watermark classification label for the data sample.

Systems, methods, and computer program products for determining an attack on a neural network. A data sample is received at a first classifier neural network and at a watermark classifier neural network, wherein the first classifier neural network is trained using a first dataset and a watermark dataset. The first classifier neural network determines a classification label for the data sample. A watermark classifier neural network determines a watermark classification label for the data sample. A data sample is determined as an adversarial data sample based on the classification label for the data sample and the watermark classification label for the data sample.

Systems and methods are described for obfuscating variants of content segments. Variants of content segments can be used to encode an identifying sequence in a transmission of content. The variants of the content segments can each include one or more marked frames and one or more unmarked frames. Variations can be introduced into the unmarked frames for each of the variants of the content segments.

The present disclosure a method of providing identification code insertion service for tracking a duplicated image, which is performed by a server, including: (a) receiving an image from a user terminal; (b) converting the received image to black and white, and selecting a plurality of insertion regions in the converted image; (c) transforming an image of at least one of a plurality of insertion regions selected at random; and (d) mapping an identification code and image information included in the transformed image of the insertion region, storing the identification code and the image information in a database, and providing the image in which the identification code is inserted to the user terminal.

A device comprises an image handling module for protecting an image of a scene captured by a camera. The image handling module is configured detect, in association with the image, a watermark signal having been embedded in light illuminating the scene at a respective geographic location. Further, the image handling module is configured to lookup the detected watermark signal in a privacy database, and based thereon to selectively inhibit use of the image.

A client forensic watermarking device, system, and method. A forensic watermarking device capable of communicating with a content server selecting a watermark mask area in which a watermark mask is displayed from video content and storing watermark area information about the watermark mask area in a storage unit according to the present disclosure may provide: a downloading unit requesting the video content to be played from the content server and receiving the video content and the watermark area information from the content server; a watermark mask generation unit outputting the watermark mask using the watermark area information inputted from the downloading unit; and an overlay unit superimposing the watermark mask inputted from the watermark mask generation unit on the watermark mask area of the video content inputted from the downloading unit, thereby enabling a client to display a forensic watermark so as to deal with a collusion attack.

Systems, methods, and computer program products for determining an attack on a neural network. A data sample is received at a first classifier neural network and at a watermark classifier neural network, wherein the first classifier neural network is trained using a first dataset and a watermark dataset. The first classifier neural network determines a classification label for the data sample. A watermark classifier neural network determines a watermark classification label for the data sample. A data sample is determined as an adversarial data sample based on the classification label for the data sample and the watermark classification label for the data sample.

Provided is an apparatus for embedding a plurality of forensic marks comprising: a pre-processing unit configured to: embed a watermark 0 symbol in each section content of an original content and store 0-section contents as a 0-content file and embed a watermark 1 symbol in each section content of the original content and store 1-section contents as a 1-content file; and embed random information in at least one section content among the 0-section contents and the 1-section contents and store random information section contents as a random information content file; and a distribution unit configured to: select corresponding section contents of the 0-content file and the 1-content file using predetermined information that is based on metadata; if a random information section content is present, select the random information section content instead of a 0-section content or a 1-section content; and output the selected random information section content as a distribution content.

Systems and methods are described for obfuscating variants of content segments. Variants of content segments can be used to encode an identifying sequence in a transmission of content. The variants of the content segments can each include one or more marked frames and one or more unmarked frames. Variations can be introduced into the unmarked frames for each of the variants of the content segments.

Systems and methods are described for obfuscating variants of content segments. Variants of content segments can be used to encode an identifying sequence in a transmission of content. The variants of the content segments can each include one or more marked frames and one or more unmarked frames. Variations can be introduced into the unmarked frames for each of the variants of the content segments.