Covert and robust mark for media identification
View Patent ↗Systems and methods are described for repeatedly embedding information in media that is recoverable by emphasizing subtle variations between the media. In many embodiments perceptual models are used to identify portions of the media in which the embedded information is unlikely to be apparent to a human observer. One embodiment of the invention includes rendering the information in the form of a media, selecting a plurality of portions of media from the carrier media, combining at least part of the rendered media with each of the selected portions of the carrier media to form modified media portions, where the modifications to the modified media portions are not apparent to a human observer and creating a modified carrier medium using the modified media portions.
1. A method of imperceptibly and robustly embedding information as a human recognizable image in a video sequence, comprising:
rendering the information as a human recognizable image, where the information is visually represented within the human recognizable image;
obtaining a plurality of video frames from the video sequence;
directly combining at least part of the rendered human recognizable image with each of the plurality of video frames to form modified video frames, where the parts of the rendered human recognizable image are embedded in corresponding embedding positions across the plurality of video frames and can be used to recreate the rendered human recognizable image by combining the plurality of video frames, and the modifications to the video frames are not apparent to a human observer; and
thereby creating a modified video sequence containing the human recognizable image, using the modified video frames.
2. The method of claim 1 further comprising varying the geometry of the rendered human recognizable image.
3. The method of claim 1 wherein combining parts of the rendered human recognizable image with each of the plurality of video frames includes modifying the rendered human recognizable image according to a perceptual model.
4. The method of claim 1 wherein the portions of the rendered human recognizable image that are combined with video frames are selected pseudo randomly.
5. The method of claim 1 wherein a lens is used to combine at least part of the human recognizable image with each of the plurality of video frames.
6. The method of claim 1 wherein the information comprises metadata related to the media.
7. The method of claim 1 wherein the information identifies the time and location of video playback.
8. The method of claim 1 wherein the information identifies the copyright owner or recipient of the media.
9. The method of claim 1 wherein the information includes a reference to a database.
10. A method of uncovering information represented within a human recognizable image robustly and imperceptibly embedded in a marked video sequence, where the marked video sequence is generated by combining at least part of the human recognizable image with each of a plurality of frames of a video sequence in corresponding embedding locations, comprising:
using video frames from the marked video sequence, where parts of the human recognizable image are imperceptibly present in a plurality of the video frames; and
combining the video frames in a predetermined manner that results in an accumulated image in which the parts of the human recognizable image present in the video frames are emphasized, where the combining of video frames in a predetermined manner proceeds until the human recognizable image is sufficiently emphasized in the accumulated image as to be recognizable.
11. The method of claim 10 wherein combining the video frames in a predetermined manner comprises averaging together the video frames.
12. The method of claim 10 , wherein the human recognizable information is recognized by a device after the combination of the video frames results in an accumulated image in which the human recognizable image is recognizable.
13. The method of claim 10 , wherein combining the video frames in a predetermined manner further comprises selecting a predetermined portion of each video frame.
14. The method of claim 10 , wherein the video sequence is degraded.
15. The method of claim 10 , further comprising selecting a plurality of video frames from the video sequence.
16. The method of claim 10 , wherein combining the video frames in a predetermined manner further comprises emphasizing at least one predetermined aspect of the video frames.
17. The method of claim 10 , wherein the plurality of video frames are of different images.
18. The method of claim 1 , wherein obtaining a plurality of video frames from the video sequence comprises selecting a plurality of video frames from the video sequence.
19. The method of claim 16 , wherein emphasizing at least one predetermined aspect of the video frames comprises applying a high pass filter to the video frames.