Image analysis – Applications
Reexamination Certificate
1997-12-30
2001-03-27
Au, Amelia (Department: 2623)
Image analysis
Applications
C382S250000, C713S176000
Reexamination Certificate
active
06208745
ABSTRACT:
The invention relates to image processing systems and, more particularly, the invention relates to a method and apparatus for performing digital watermarkingofa digital image sequence.
BACKGROUND OF THE DISCLOSURE
With the advent of digitization of images, digital image distribution and digital video availability, copyright protection of such digital imagery has become a substantial issue for image publishers and authors. One technique used to identify digital video ownership is a digital “watermark” that is embedded into an image sequence. Such watermarks must be secure and robust to intentional corruption and to compression processing, not be unreasonably complex to embed and extract, and be compatible and interoperable with conventional image processing systems. For authentication applications, the watermark is generally invisible to a viewer of a decoded image sequence. However, in some applications, it is desirable to produce a visible watermark that can be removed by an authorized image decoder and that can not be removed by an unauthorized decoder.
Various digital watermarking techniques have been attempted for both still and video images with varying levels of success. The use of spread-spectrum techniques are disclosed in Hartung et al., “Digital Watermarking of Raw and Compressed Video”, Systems for Video Communication, October 1996, pp. 205-213 and Hartung et al., “Watermarking of MPEG-2 Encoded Video Without Decoding and Reencoding”, Proceedings of SPIE 3020, Multimedia Computing and Networking 97 (MMCN 97), February 1997. The first of these papers discloses a technique that spreads the energy of a watermark image throughout a video sequence to be “watermarked” using a pseudo-noise signal. Once the pseudo-noise signal has been embedded into the video sequence, the system encodes the video sequence containing the watermark. In this manner, the watermarking is accomplished in the pixel domain. As such, any coding losses in the video coding process apply to the watermark as well as the images in the video sequence. At the decoder, the spread watermark is correlated and extracted from the video sequence. The use of a spread-spectrum technique requires substantial synchronization and signal processing hardware at both the encoder and decoder to facilitate recovery of the watermark and the video. As such, pixel domain watermark processes are generally avoided.
The second of the aforementioned papers disclose a bitstream domain watermarking technique where the “block” of an image frame within the video sequence is coded and then combined with a coded watermark signal. Specifically, a block is generally an 8×8 pixel portion of an image frame. The block is coded using a discrete cosine transform (DCT) to form a coded block. A watermark image is similarly divided into blocks and DCT coded. The DCT coefficients representing the coded watermark block and the coded image block are then added together to form a combined block. The combined block is quantized and error coded. Thereafter, a selection process is performed to transmit only the “watermarked” coefficients that will not increase the bit rate necessary to transmit the encoded video sequence. This selection process compares the number of bits required to encode coefficients of the combined block with the number of bits required to encode coefficients of the image block only. If the number of combined block bits is greater than the number of image block bits, the system transmits the image block bits; otherwise, the combined block is transmitted. Clearly, such a selection process eliminates some of the watermark information and thus distorts the watermark. In some extreme cases, the watermark information may not be transmitted at all or such a small amount may be transmitted that the watermark is rendered useless. Also, due to the randomness of transmitting or not transmitting each bit, this technique is restricted to spread spectrum or other highly redundant watermarking techniques. The redundancy may be used by an attacker to locate and remove the watermark.
Therefore, a need exists in the art for a watermarking technique that does not remove any amount of watermark signal by inserting a watermark into the encoded bitstream representation of an image sequence.
SUMMARY OF THE INVENTION
The disadvantages associated with the prior art are overcome by the present invention of a method and apparatus that inserts watermark information directly into an encoded video bitstream. The invention identifies certain blocks or macroblocks in an encoded video bitstream and inserts the watermark information directly into the bitstream such that these certain blocks or macroblocks are replaced with a block or macroblock containing watermark information. In its broadest sense, the invention may replace video information in any macroblock or block within a macroblock with a macroblock or block containing watermark information. One form of macroblock that can be replaced with watermark information is a “skipped” macroblock. So-called skipped macroblocks are macroblocks of video information that are deemed inconsequential to the decoding process and, as such, are identified by the encoder as macroblocks that are not to be encoded and transmitted, i.e., skipped. Additionally, macroblocks may contain blocks of information that may be “skipped” and, as such, although a macroblock itself is not skipped, a block within the macroblock may be skipped. Consequently, the invention may insert watermark information into such a skipped block.
In a block-based encoding technique such as MPEG encoding, the video is encoded as defined groups of pictures, where each group of pictures contains particular types of frames known as I, P and B frames. If the watermark information is placed in blocks within B-frames only, then the watermark will generally be invisible. B-frames are not used to predict other frames during decoding; therefore, the watermark will only appear on the screen for a fraction of a second. If, however, the watermark information is placed in an I- or P-frame, where the frame information is used to predict other frames during decoding, the watermark information will be visible in the decoded video unless the watermark is removed from the bitstream by an “authorized” decoder. To further adjust the visibility of the watermark, the amplitude of the discrete cosine transform coefficients that comprise a watermark is adjusted as well as the quantization scale that is used to quantize the watermark DCT coefficients.
REFERENCES:
patent: 5530759 (1996-06-01), Braudaway et al.
patent: 5568570 (1996-10-01), Rabbani
patent: 5636292 (1997-06-01), Rhoads
patent: 5646997 (1997-07-01), Barton
patent: 5659726 (1997-08-01), Sandford, II et al.
patent: 5661574 (1997-08-01), Kawana
patent: 5664018 (1997-09-01), Leighton
patent: 5778102 (1998-07-01), Sandford, II et al.
patent: 5825892 (1998-10-01), Braudaway et al.
patent: 5848155 (1998-12-01), Cox
patent: 5862260 (1999-01-01), Rhoads
patent: 5915027 (1999-06-01), Cox et al.
patent: 5933798 (1999-08-01), Linnartz
patent: 0 805 600 A2 (1997-11-01), None
patent: 0 840 513 A2 (1998-05-01), None
Adaptive Watermarking in the DCT Domain, Tao et al; IEEE Conference on Acoustics, Speech, and Signal Processing, ISBN: 0-8186-7919-0, vol. 4, pp. 2985-2988, 1997.*
I. Cox, J. Kilian, T. Leighton, T. Shamoon “Secure Spread Spectrum Watermarking for Multimedia”, NEC Research Institute, Technical Report 95-10.
F. Hartung, B. Girod, “Watermarking of MPEG-2 encoded video without decoding and re-encoding”, Proceedings SPIE 3020: Multimedia Computing and Networking 97 (MMCN 97) Feb. 1997.
F. Hartung, B. Girod, “Digital Watermarking of Raw and Compressed Video”, Systems for Video Communication, pp. 205-213, Oct. 1996.
Hsu Chiou-Ting et al: “Digital watermarking for video” 1997, Proceedings of the 1997 13th International Conference on Digital Signal Processing, DSP. Part 1 (of 2); Santorini, Greece Jul. 2-4 1997, Int Conf Dig Signal Process DSP; International Conference on Digital Signal Processing, DSP 1997 IEEE Piscatawa
Florencio Dinei A.
Isnardi Michael A.
Au Amelia
Burke William J.
Dastouri Mehrdad
Sarnoff Corporation
LandOfFree
Method and apparatus for imbedding a watermark into a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for imbedding a watermark into a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for imbedding a watermark into a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2480442