Embedding watermarks into compressed video data
First Claim
1. A method for adding a watermark to a compressed video bitstream comprising the steps of:
- selecting a plurality of sites within the compressed video bitstream that encode a discrete cosine transform (“
DCT”
) coefficient which is apt for modification to embed a watermark into the compressed video bitstream;
modifying the DCT coefficient for at least some of the selected plurality of sites so the modified DCT coefficient occupies a number of bits in the compressed video bitstream that does not exceed a number of bits occupied by that same DCT coefficient before modification, thereby embedding the watermark into the compressed video bitstream; and
transmitting the watermarked compressed video bitstream containing the modified DCT coefficients.
1 Assignment
0 Petitions
Accused Products
Abstract
A for method watermarking a compressed video bitstream selects for modification numerous sites within an MPEG compressed video bitstream, each of which sites includes a DCT coefficient which is apt for modification. Watermarking sites are preferably located in MPEG B frames, and the DCT coefficient to be modified preferably has a run-length of zero (0). The method then embeds the watermark by modifying the DCT coefficient for at least some of the selected sites. A list specifying the watermarking sites may be transmitted together with the compressed video bitstream to permit further watermarking by subsequent bitstream receivers. A system for detecting an unauthorized copy of a watermarked video forms a synchronized, frame-by-frame difference between a possibly unauthorized copy of the watermarked compressed video and an un-watermarked copy of the video. Data in the watermarking site list permits readily locating watermarks within the possibly unauthorized copy.
-
Citations
32 Claims
-
1. A method for adding a watermark to a compressed video bitstream comprising the steps of:
-
selecting a plurality of sites within the compressed video bitstream that encode a discrete cosine transform (“
DCT”
) coefficient which is apt for modification to embed a watermark into the compressed video bitstream;
modifying the DCT coefficient for at least some of the selected plurality of sites so the modified DCT coefficient occupies a number of bits in the compressed video bitstream that does not exceed a number of bits occupied by that same DCT coefficient before modification, thereby embedding the watermark into the compressed video bitstream; and
transmitting the watermarked compressed video bitstream containing the modified DCT coefficients. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
receiving the compressed video bitstream before selecting the plurality of sites.
-
-
7. The method for adding a watermark to a compressed video bitstream of claim 1 further comprising the step of:
generating the compressed video bitstream by compressing a received video signal concurrently with selecting the plurality of sites.
-
8. The method for adding a watermark to a compressed video bitstream of claim 1 further comprising the step of:
generating the compressed video bitstream by compressing a received video signal before selecting the plurality of sites.
-
9. The method for adding a watermark to a compressed video bitstream of claim 1 further comprising the step of:
also transmitting site data specifying sites within the compressed video bitstream which are apt for modification to embed a watermark into the compressed video bitstream.
-
10. The method for adding a watermark to a compressed video bitstream of claim 1 wherein when the modified DCT coefficient occupies a number of bits that is less than a number of bits occupied by that same DCT coefficient before modification, a number of zeros (0'"'"'s) are prefixed onto the modified DCT coefficient so a total of:
-
the number of bits in the modified DCT coefficient;
plusthe number of bits in the added zeros (0'"'"'s);
is identical the number of bits occupied by the DCT coefficient before modification.
-
-
11. A method for adding a watermark to a compressed video bitstream comprising the steps of:
-
receiving the compressed video bitstream together with site data specifying a plurality of sites within the compressed video bitstream which are apt for modification to embed a watermark into the compressed video bitstream;
modifying a DCT coefficient for at least some of the plurality of sites specified by the received site data thereby embedding the watermark into the compressed video bitstream; and
re-transmitting the watermarked compressed video bitstream containing the modified DCT coefficients. - View Dependent Claims (12, 13)
also transmitting at least some of the site data specifying sites within the compressed video bitstream which are apt for modification to embed a watermark into the compressed video bitstream.
-
-
14. A method for adding a watermark to a compressed video bitstream comprising the steps of:
-
selecting a plurality of sites within the compressed video bitstream that encode a DCT coefficient which is apt for modification to embed a watermark into the compressed video bitstream; and
transmitting both the compressed video bitstream and site data specifying the selected plurality of sites within the compressed video bitstream which are apt for modification to embed a watermark into the compressed video bitstream. - View Dependent Claims (15, 16, 17, 18)
-
-
19. A method for adding a watermark to a video signal comprising the steps of:
-
receiving the compressed video bitstream together with site data specifying a plurality of sites within the compressed video bitstream which are apt for modification to embed a watermark into the compressed video bitstream;
modifying a DCT coefficient for at least some of the plurality of sites thereby embedding the watermark into the compressed video bitstream; and
decoding the watermarked compressed video bitstream containing the modified DCT coefficients to produce therefrom a video signal that is suitable for display on a television set, and that has the watermark embedded therein. - View Dependent Claims (20)
-
-
21. A method for adding a watermark to a video signal comprising the steps of:
-
receiving the compressed video bitstream;
selecting a plurality of sites within the compressed video bitstream that encode a DCT coefficient which is apt for modification to embed a watermark into the compressed video bitstream;
modifying the DCT coefficient for at least some of the selected plurality of sites so the modified DCT coefficient occupies a number of bits in the compressed video bitstream that does not exceed a number of bits occupied by that same DCT coefficient before modification, thereby embedding the watermark into the compressed video bitstream; and
decoding the compressed video bitstream containing the modified DCT coefficients to produce therefrom a video signal that is suitable for display on a television set, and that has the watermark embedded therein. - View Dependent Claims (22, 23, 24, 25, 26)
the number of bits in the modified DCT coefficient;
plusthe number of bits in the added zeros (0'"'"'s);
is identical the number of bits occupied by the DCT coefficient before modification.
-
-
27. A system for detecting an unauthorized copy of a video by identifying a watermark that has been embedded into a compressed video bitstream produced from an un-watermarked video by changing values of selected DCT coefficients at watermarking sites in the compressed video bitstream, the system comprising:
-
an original video input for receiving a video signal of the un-watermarked video;
a copy video input for receiving a video signal of the copy of the video which may possibly include the watermark;
a frame differencer that receives from said original video input and from said copy video input the video signals received thereat, said frame differencer producing a synchronized, frame-by-frame difference between the video signal received by said original video input and the video signal received by said copy video input; and
watermark identification means that receives the frame-by-frame difference produced by said frame differencer and also receives a site list containing data which specifies characteristics of sites at which the watermark may be embedded into the video signal received by said copy video input, said watermark identification means permitting detection of the watermark embedded into the video signal received by said copy video input. - View Dependent Claims (28, 29, 30, 31, 32)
said original video input receives an analog video signal; and
the system further includes a video digitizer for digitizing the analog video signal received by said original video input, and for supplying the digitized video signal thus produced to said frame differencer.
-
-
29. The system of claim 27 wherein:
-
said copy video input receives an analog video signal; and
the system further includes a video digitizer for digitizing the analog video signal received by said copy video input, and for supplying the digitized video signal thus produced to said frame differencer.
-
-
30. The system of claim 27 wherein said watermark identification means includes:
-
a digital-to-analog converter for receiving the frame-by-frame difference produced by said frame differencer and for converting the received frame-by-frame difference into an analog video signal; and
a video monitor for receiving the analog video signal produced by said digital-to-analog converter and for visually displaying watermarking sites.
-
-
31. The system of claim 27 wherein said watermark identification means includes a frame analyzer for receiving the frame-by-frame difference produced by said frame differencer and for automatically determining if a watermark occurs at a watermarking site.
-
32. The system of claim 31 wherein in determining if a watermark occurs at a watermarking site said frame analyzer computes a signal-to-noise ratio (“
- SNR”
) between;the DCT coefficient at the watermarking site in a discrete cosine transform (“
DCT”
) computed from the frame-by-frame difference produced by said frame differencer; and
other coefficients of the DCT computed from the frame-by-frame difference.
- SNR”
Specification