Image analysis – Applications
Reexamination Certificate
1999-08-04
2004-08-17
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
Reexamination Certificate
active
06778678
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to image processing techniques, and more particularly to techniques for processing images to incorporate digital watermarking information.
BACKGROUND OF THE INVENTION
Digital watermarking techniques are used to protect electronic data from unauthorized copying or distribution. Unlike a traditional visible watermark used on paper, a digital image watermark is generally designed so as not to alter the perceived quality of the electronic content, while also being robust to attacks. For example, in the case of image data, typical signal processing operations, such as linear and nonlinear filtering, cropping, rescaling, noise removal, lossy compression, etc., should ideally be configured such that if any of these operations result in alteration or suppression of the inserted watermark, then the resulting image must be so severely degraded to render it worthless. However, it is equally important that the digital watermark not alter the perceived visual quality of the image. It is therefore clear that from a signal processing viewpoint, the two basic requirements for an effective watermarking technique, i.e., robustness and transparency, conflict with each other.
Digital watermarking applications can generally be grouped into two main categories: source-based applications and destination-based applications. Digital watermarks in source-based applications are typically used for purposes of ownership identification and tamper detection. A unique watermark signal is hidden in all copies of a particular image, prior to their distribution. Examination of the particular watermark signal hidden in a given image can then be used to determine the originator of the image, and whether parts of the image have been tampered with, e.g., if the picture in a photo identification has had the face replaced, etc. Furthermore, digital watermarks can be used to embed application-dependent information, not necessarily dealing with security issues, that can be maintained even when the image is transferred across different media such as disk, D1 tape, high-quality printouts, etc.
Digital watermarks in destination-based applications are typically used for tracing purposes. In such applications, a distinct watermark signal that uniquely identifies a particular copy of the image is hidden in that copy, prior to its distribution, and acts as a “serial number” for the image. Then, in the event that multiple unauthorized copies of a given image are detected, retrieval of that serial number from one of the copies of the image can identify the particular user whose image was utilized to create the unauthorized copies.
It is known that spread-spectrum communication techniques, as described in, e.g., R. Blahut, “Digital Transmission of Information,” Addison Wesley Publishing Company, 1990, can also be applied to increase the robustness of digital watermarks. In spread-spectrum communication systems, an information-bearing narrowband signal is converted into a wideband signal prior to transmission, by modulating the information waveform with a wideband noise-like waveform that is unknown to a jammer. As a result of this bandwidth expansion, within any narrow spectral band, the total amount of energy from the information signal is small. However, by appropriately combining all these weak narrowband signals at the demodulator, the original information signal is recovered. Hence a jammer, unaware of the shape of the wideband carrier, is forced to spread its available power over a much larger bandwidth, thus reducing its effectiveness.
The application of the above-described spread-spectrum communication techniques to digital watermarking is described in, e.g., I. Cox, J. Killian, T. Leighton, and T. Shamoon, “Secure Spread Spectrum Watermarking for Multimedia,” Technical Report 95-10, NEC Research Institute, 1995. In this approach, robustness and transparency are ensured by introducing many small changes into the most perceptually-significant image components. Since during the watermark extraction process the location and value of these changes are known, it is possible to concentrate the information of all the small changes to come up with a robust decision on the presence or absence of a particular digital watermark. Furthermore, in order to destroy such a watermark, a substantial amount of noise would be required in all the perceptually-significant components, thereby drastically reducing the perceived image quality.
These and other conventional digital watermarking techniques may make use of models of the human visual system. Recently, visual models have been developed specifically for the performance evaluation of lossy image compression algorithms, e.g., A. Watson, G. Yang, J. Solomon, and J. Villasenor, “Visibility of Wavelet Quantization Noise,” IEEE Transactions on Image Processing, 6(8), pp.1164-1175, August 1997. One common paradigm for perceptual image coding is based on deriving an image dependent mask containing a set of just noticeable difference (JND) thresholds used to compute perceptually-based quantizers. These models, originally designed for perceptual coding applications, are also well suited for watermarking. For example, the JND thresholds can be used as upper bounds on watermark intensity levels. Hence, a criterion is available to address simultaneously the conflicting goals of robustness and transparency: a watermark can be made maximally strong, subject to an invisibility constraint determined from the JND thresholds. An effective watermarking technique based on these principles is described in C. Podilchuk and W. Zeng, “Image Adaptive Watermarking Using Visual Models,” IEEE Journal on Selected Areas in Communications, 16(4), May 1998.
Other conventional digital image watermarking techniques are described in, e.g., J. O. Ruanaidh, W. Dowling, and F. Boland, “Watermarking Digital Images for Copyright Protection,” IEEE Proceedings on Vision, Image and Signal Processing, 143(4), pp. 250-256, August 1996, and J. Smith and B. Comiskey, “Modulation and Information Hiding in Images,” Lecture Notes in Computer Science (1174), Springer-Verlag, August 1996. A significant problem with these and other conventional techniques is that they fail to address adequately the issue of how many watermarks can be reliably encoded. For example, because the approach in the J. Smith and B. Comiskey reference fails to differentiate between image data and impairments introduced by a jammer, the number of different watermarks that can be reliably distinguished is significantly reduced.
SUMMARY OF THE INVENTION
In accordance with the invention, digital watermark information is inserted into an image by first separating the image into components, e.g., discrete cosine transform (DCT) blocks or image subbands, and then associating one or more bits of the digital watermark information with each of the components. In an illustrative embodiment of the invention, a single bit of digital watermark information is associated with each of the components by modulating the components with selected waveforms representative of the corresponding digital watermark information bits. For example, the selected waveforms may comprise a pair of n-bit vectors having a zero mean and an identity covariance matrix, with one of the vectors representing a binary one, and the other representing a binary zero. The digital watermark information may include a total of B bits of information for representing a particular watermark, such that M=2
B
distinct watermarks can be generated using the B information bits. A visual model may be used to determine a particular subset of the image components to be associated with one or more bits of the digital watermark information, so as to ensure that modification or deletion of the watermark information will render the resulting image unusable.
In another possible embodiment of the invention, the digital watermark information bits may be coded, e.g., using a repetition code, linear block code or convolutional code, to form channel bits, such tha
Podilchuk Christine Irene
Servetto Sergio Daniel
Johns Andrew W.
Lucent Technologies - Inc.
Nakhjavan Shervin
LandOfFree
High-capacity digital image watermarking based on waveform... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with High-capacity digital image watermarking based on waveform..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and High-capacity digital image watermarking based on waveform... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3356646