Methods for surveying dissemination of proprietary empirical data

Abstract

An automated system checks networked computers, such as computers on the internet, for watermarked audio, video, or image data. A report listing the location of such audio, video or image data is generated, and provided to the proprietor(s) thereof identified by the watermark information.

Description

RELATED APPLICATION DATA

This application is a continuation in part of the following copending applications: PCT/US96/06618 (May 7, 1996), Ser. No. 08/637,531 (filed Apr. 25, 1996) allowed, Ser. No. 08/534,005 (filed Sep. 25, 1995) allowed, Ser. No. 10
Here, n and m are simple indexing values on rows and columns of the image ranging from 0 to 3999. Sqrt is the square root. V is the DN of a given indexed pixel on the original digital image. The < >brackets around the RMS noise merely indicates that this is an expected average value, where it is clear that each and every pixel will have a random error individually. Thus, for a pixel value having 1200 as a digital number or “brightness value”, we find that its expected rms noise value is sqrt(1204)=34.70, which is quite close to 34.64, the square root of 1200.

We furthermore realize that the square root of the innate brightness value of a pixel is not precisely what the eye perceives as a minimum objectionable noise, thus we come up with the formula:
<RMS Addable Noise_n,m>=X*sqrt(4+(V_n,m−30)⁻Y)   (2)
Where X and Y have been added as empirical parameters which we will adjust, and “addable” noise refers to our acceptable perceived noise level from the definitions above. We now intend to experiment with what exact value of X and Y we can choose, but we will do so at the same time that we are performing the next steps in the process.

The next step in our process is to choose N of our N-bit identification word. We decide that a 16 bit main identification value with its 65536 possible values will be sufficiently large to identify the image as ours, and that we will be directly selling no more than 128 copies of the image which we wish to track, giving 7 bits plus an eighth bit for an odd/even adding of the first 7 bits (i.e. an error checking bit on the first seven). The total bits required now are at 4 bits for the 0101 calibration sequence, 16 for the main identification, 8 for the version, and we now throw in another 4 as a further error checking value on the first 28 bits, giving 32 bits as N. The final 4 bits can use one of many industry standard error checking methods to choose its four values.

We now randomly determine the 16 bit main identification number, finding for example, 1101 0001 1001 1110; our first versions of the original sold will have all 0's as the version identifier, and the error checking bits will fall out where they may. We now have our unique 32 bit identification word which we will embed on the original digital image.

To do this, we generate 32 independent random 4000 by 4000 encoding images for each bit of our 32 bit identification word. The manner of generating these random images is revealing. There are numerous ways to generate these. By far the simplest is to turn up the gain on the same scanner that was used to scan in the original photograph, only this time placing a pure black image as the input, then scanning this 32 times. The only drawback to this technique is that it does require a large amount of memory and that “fixed pattern” noise will be part of each independent “noise image.” But, the fixed pattern noise can be removed via normal “dark frame” subtraction techniques. Assume that we set the absolute black average value at digital number ‘100,’ and that rather than finding a 2 DN rms noise as we did in the normal gain setting, we now find an rms noise of 10 DN about each and every pixel's mean value.

We next apply a mid-spatial-frequency bandpass filter (spatial convolution) to each and every independent random image, essentially removing the very high and the very low spatial frequencies from them. We remove the very low frequencies because simple real-world error sources like geometrical warping, splotches on scanners, mis-registrations, and the like will exhibit themselves most at lower frequencies also, and so we want to concentrate our identification signal at higher spatial frequencies in order to avoid these types of corruptions. Likewise, we remove the higher frequencies because multiple generation copies of a given image, as well as compression-decompression transformations, tend to wipe out higher frequencies anyway, so there is no point in placing too much identification signal into these frequencies if they will be the ones most prone to being attenuated. Therefore, our new filtered independent noise images will be dominated by mid-spatial frequencies. On a practical note, since we are using 12-bit values on our scanner and we have removed the DC value effectively and our new rms noise will be slightly less than 10 digital numbers, it is useful to boil this down to a 6-bit value ranging from −32 through 0 to 31 as the resultant random image.

Next we add all of the random images together which have a ‘1’ in their corresponding bit value of the 32-bit identification word, accumulating the result in a 16-bit signed integer image. This is the unattenuated and un-scaled version of the composite embedded signal.

Next we experiment visually with adding the composite embedded signal to the original digital image, through varying the X and Y parameters of equation 2. In formula, we visually iterate to both maximize X and to find the appropriate Y in the following.
V_dist;n,m=V_orig;n,m+
where dist refers to the candidate distributable image, i.e. we are visually iterating to find what X and Y will give us an acceptable image; orig refers to the pixel value of the original image; and comp refers to the pixel value of the composite image. The n's and m's still index rows and columns of the image and indicate that this operation is done on all 4000 by 4000 pixels. The symbol V is the DN of a given pixel and a given image.

As an arbitrary assumption, now, we assume that our visual experimentation has found that the value of X=0.025 and Y=0.6 are acceptable values when comparing the original image with the candidate distributable image. This is to say, the distributable image with the “extra noise” is acceptably close to the original in an aesthetic sense. Note that since our individual random images had a random rms noise value around 10 DN, and that adding approximately 16 of these images together will increase the composite noise to around 40 DN, the X multiplication value of 0.025 will bring the added rms noise back to around 1 DN, or half the amplitude of our innate noise on the original. This is roughly a 1 dB gain in noise at the dark pixel values and correspondingly more at the brighter values modified by the Y value of 0.6.

So with these two values of X and Y, we now have constructed our first versions of a distributable copy of the original. Other versions will merely create a new composite signal and possibly change the X slightly if deemed necessary. We now lock up the original digital image along with the 32-bit identification word for each version, and the 32 independent random 4-bit images, waiting for our first case of a suspected piracy of our original. Storage wise, this is about 14 Megabytes for the original image and 32*0.5 bytes*16 million=−256 Megabytes for the random individual encoded images. This is quite acceptable for a single valuable image. Some storage economy can be gained by simple lossless compression.

Finding a Suspected Piracy of our Image

We sell our image and several months later find our two heads of state in the exact poses we sold them in, seemingly cut and lifted out of our image and placed into another stylized background scene. This new “suspect” image is being printed in 100,000 copies of a given magazine issue let us say. We now go about determining if a portion of our original image has indeed been used in an unauthorized manner. FIG. 3 summarizes the details.

The first step is to take an issue of the magazine, cut out the page with the image on it, then carefully but not too carefully cut out the two figures from the background image using ordinary scissors. If possible, we will cut out only one connected piece rather than the two figures separately. We paste this onto a black background and scan this into a digital form. Next we electronically flag or mask out the black background, which is easy to do by visual inspection.

We now procure the original digital image from our secured place along with the 32-bit identification word and the 32 individual embedded images. We place the original digital image onto our computer screen using standard image manipulation software, and we roughly cut along the same borders as our masked area of the suspect image, masking this image at the same time in roughly the same manner. The word ‘roughly’ is used since an exact cutting is not needed, it merely aids the identification statistics to get it reasonably close.

Next we rescale the masked suspect image to roughly match the size of our masked original digital image, that is, we digitally scale up or down the suspect image and roughly overlay it on the original image. Once we have performed this rough registration, we then throw the two images into an automated scaling and registration program. The program performs a search on the three parameters of x position, y position, and spatial scale, with the figure of merit being the mean squared error between the two images given any given scale variable and x and y offset. This is a fairly standard image processing methodology. Typically this would be done using generally smooth interpolation techniques and done to sub-pixel accuracy. The search method can be one of many, where the simplex method is a typical one.

Once the optimal scaling and x-y position variables are found, next comes another search on optimizing the black level, brightness gain, and gamma of the two images. Again, the figure of merit to be used is mean squared error, and again the simplex or other search methodologies can be used to optimize the three variables. After these three variables are optimized, we apply their corrections to the suspect image and align it to exactly the pixel spacing and masking of the original digital image and its mask. We can now call this the standard mask.

The next step is to subtract the original digital image from the newly normalized suspect image only within the standard mask region. This new image is called the difference image.

Then we step through all 32 individual random embedded images, doing a local cross-correlation between the masked difference image and the masked individual embedded image. ‘Local’ refers to the idea that one need only start correlating over an offset region of ±1 pixels of offset between the nominal registration points of the two images found during the search procedures above. The peak correlation should be very close to the nominal registration point of 0,0 offset, and we can add the 3 by 3 correlation values together to give one grand correlation value for each of the 32 individual bits of our 32-bit identification word.

After doing this for all 32 bit places and their corresponding random images, we have a quasi-floating point sequence of 32 values. The first four values represent our calibration signal of 0101. We now take the mean of the first and third floating point value and call this floating point value ‘0,’ and we take the mean of the second and the fourth value and call this floating point value ‘1.’ We then step through all remaining 28 bit values and assign either a ‘0’ or a ‘1’ based simply on which mean value they are closer to. Stated simply, if the suspect image is indeed a copy of our original, the embedded 32-bit resulting code should match that of our records, and if it is not a copy, we should get general randomness. The third and the fourth possibilities of 3) Is a copy but doesn't match identification number and 4) isn't a copy but does match are, in the case of 3), possible if the signal to noise ratio of the process has plummeted, i.e. the ‘suspect image’ is truly a very poor copy of the original, and in the case of 4) is basically one chance in four billion since we were using a 32-bit identification number. If we are truly worried about 4), we can just have a second independent lab perform their own tests on a different issue of the same magazine. Finally, checking the error-check bits against what the values give is one final and possibly overkill check on the whole process. In situations where signal to noise is a possible problem, these error checking bits might be eliminated without too much harm.

Benefits

Now that a full description of the first embodiment has been described via a detailed example, it is appropriate to point out the rationale of some of the process steps and their benefits.

The ultimate benefits of the foregoing process are that obtaining an identification number is fully independent of the manners and methods of preparing the difference image. That is to say, the manners of preparing the difference image, such as cutting, registering, scaling, etcetera, cannot increase the odds of finding an identification number when none exists; it only helps the signal-to-noise ratio of the identification process when a true identification number is present. Methods of preparing images for identification can be different from each other even, providing the possibility for multiple independent methodologies for making a match.

The ability to obtain a match even on sub-sets of the original signal or image is a key point in today's information-rich world. Cutting and pasting both images and sound clips is becoming more common, allowing such an embodiment to be used in detecting a copy even when original material has been thus corrupted. Finally, the signal to noise ratio of matching should begin to become difficult only when the copy material itself has been significantly altered either by noise or by significant distortion; both of these also will affect that copy's commercial value, so that trying to thwart the system can only be done at the expense of a huge decrease in commercial value.

An early conception of this technology was the case where only a single “snowy image” or random signal was added to an original image, i.e. the case where N=1. “Decoding” this signal would involve a subsequent mathematical analysis using (generally statistical) algorithms to make a judgment on the presence or absence of this signal. The reason this approach was abandoned as the preferred embodiment was that there was an inherent gray area in the certainty of detecting the presence or absence of the signal. By moving onward to a multitude of bit planes, i.e. N>1, combined with simple pre-defined algorithms prescribing the manner of choosing between a “0” and a “1”, the certainty question moved from the realm of expert statistical analysis into the realm of guessing a random binary event such as a coin flip. This is seen as a powerful feature relative to the intuitive acceptance of this technology in both the courtroom and the marketplace. The analogy which summarizes the inventor's thoughts on this whole question is as follows: The search for a single identification signal amounts to calling a coin flip only once, and relying on arcane experts to make the call; whereas the N>1 embodiment relies on the broadly intuitive principle of correctly calling a coin flip N times in a row. This situation is greatly exacerbated, i.e. the problem of “interpretation” of the presence of a single signal, when images and sound clips get smaller and smaller in extent.

Another important reason that the N>1 case is preferred over the N=1 embodiment is that in the N=1 case, the manner in which a suspect image is prepared and manipulated has a direct bearing on the likelihood of making a positive identification. Thus, the manner with which an expert makes an identification determination becomes an integral part of that determination. The existence of a multitude of mathematical and statistical approaches to making this determination leave open the possibility that some tests might make positive identifications while others might make negative determinations, inviting further arcane debate about the relative merits of the various identification approaches. The N>1 embodiment avoids this further gray area by presenting a method where no amount of pre-processing of a signal—other than pre-processing which surreptitiously uses knowledge of the private code signals—can increase the likelihood of “calling the coin flip N times in a row.”

The fullest expression of the present system will come when it becomes an industry standard and numerous independent groups set up with their own means or ‘in-house’ brand of applying embedded identification numbers and in their decipherment. Numerous independent group identification will further enhance the ultimate objectivity of the method, thereby enhancing its appeal as an industry standard.

Use of True Polarity in Creating the Composite Embedded Code Signal

The foregoing discussion made use of the 0 and 1 formalism of binary technology to accomplish its ends. Specifically, the 0's and 1's of the N-bit identification word directly multiplied their corresponding individual embedded code signal to form the composite embedded code signal (step 8, FIG. 2). This approach certainly has its conceptual simplicity, but the multiplication of an embedded code signal by 0 along with the storage of that embedded code contains a kind of inefficiency.

It is preferred to maintain the formalism of the 0 and 1 nature of the N-bit identification word, but to have the 0's of the word induce a subtraction of their corresponding embedded code signal. Thus, in step 8 of FIG. 2, rather than only ‘adding’ the individual embedded code signals which correspond to a ‘1’ in the N-bit identification word, we will also ‘subtract’ the individual embedded code signals which correspond to a ‘0’ in the N-bit identification word.

At first glance this seems to add more apparent noise to the final composite signal. But it also increases the energy-wise separation of the 0's from the 1's, and thus the ‘gain’ which is applied in step 10. FIG. 2 can be correspondingly lower.

We can refer to this improvement as the use of true polarity. The main advantage of this improvement can largely be summarized as ‘informational efficiency.’

‘Perceptual Orthogonality’ of the Individual Embedded Code Signals

The foregoing discussion contemplates the use of generally random noise-like signals as the individual embedded code signals. This is perhaps the simplest form of signal to generate. However, there is a form of informational optimization which can be applied to the set of the individual embedded signals, which the applicant describes under the rubric ‘perceptual orthogonality.’ This term is loosely based on the mathematical concept of the orthogonality of vectors, with the current additional requirement that this orthogonality should maximize the signal energy of the identification information while maintaining it below some perceptibility threshold. Put another way, the embedded code signals need not necessarily be random in nature.

Use and Improvements of the First Embodiment in the Field of Emulsion-Based Photography

The foregoing discussion outlined techniques that are applicable to photographic materials. The following section explores the details of this area further and discloses certain improvements which lend themselves to a broad range of applications.

The first area to be discussed involves the pre-application or pre-exposing of a serial number onto traditional photographic products, such as negative film, print paper, transparencies, etc. In general, this is a way to embed a priori unique serial numbers (and by implication, ownership and tracking information) into photographic material. The serial numbers themselves would be a permanent part of the normally exposed picture, as opposed to being relegated to the margins or stamped on the back of a printed photograph, which all require separate locations and separate methods of copying. The ‘serial number’ as it is called here is generally synonymous with the N-bit identification word, only now we are using a more common industrial terminology.

In FIG. 2, step 11, the disclosure calls for the storage of the “original [image]” along with code images. Then in FIG. 3, step 9, it directs that the original be subtracted from the suspect image, thereby leaving the possible identification codes plus whatever noise and corruption has accumulated. Therefore, the previous disclosure made the tacit assumption that there exists an original without the composite embedded signals.

Now in the case of selling print paper and other duplication film products, this will still be the case, i.e., an “original” without the embedded codes will indeed exist and the basic methodology of the first embodiment can be employed. The original film serves perfectly well as an ‘unencoded original.’

However, in the case where pre-exposed negative film is used, the composite embedded signal pre-exists on the original film and thus there will never be an “original” separate from the pre-embedded signal. It is this latter case, therefore, which will be examined a bit more closely, along with observations on how to best use the principles discussed above (the former cases adhering to the previously outlined methods).

The clearest point of departure for the case of pre-numbered negative film, i.e. negative film which has had each and every frame pre-exposed with a very faint and unique composite embedded signal, comes at step 9 of FIG. 3 as previously noted. There are certainly other differences as well, but they are mostly logistical in nature, such as how and when to embed the signals on the film, how to store the code numbers and serial number, etc. Obviously the pre-exposing of film would involve a major change to the general mass production process of creating and packaging film.

FIG. 4 has a schematic outlining one potential post-hoc mechanism for pre-exposing film. ‘Post-hoc’ refers to applying a process after the full common manufacturing process of film has already taken place. Eventually, economies of scale may dictate placing this pre-exposing process directly into the chain of manufacturing film. Depicted in FIG. 4 is what is commonly known as a film writing system. The computer, 106, displays the composite signal produced in step 8, FIG. 2, on its phosphor screen. A given frame of film is then exposed by imaging this phosphor screen, where the exposure level is generally very faint, i.e. generally imperceptible. Clearly, the marketplace will set its own demands on how faint this should be, that is, the level of added ‘graininess’ as practitioners would put in. Each frame of film is sequentially exposed, where in general the composite image displayed on the CRT 102 is changed for each and every frame, thereby giving each frame of film a different serial number. The transfer lens 104 highlights the focal conjugate planes of a film frame and the CRT face.

Getting back to the applying the principles of the foregoing embodiment in the case of pre-exposed negative film . . . At step 9, FIG. 3, if we were to subtract the “original” with its embedded code, we would obviously be “erasing” the code as well since the code is an integral part of the original. Fortunately, remedies do exist and identifications can still be made. However, it will be a challenge to artisans who refine this embodiment to have the signal to noise ratio of the identification process in the pre-exposed negative case approach the signal to noise ratio of the case where the un-encoded original exists.

A succinct definition of the problem is in order at this point. Given a suspect picture (signal), find the embedded identification code IF a code exists at al. The problem reduces to one of finding the amplitude of each and every individual embedded code signal within the suspect picture, not only within the context of noise and corruption as was previously explained, but now also within the context of the coupling between a captured image and the codes. ‘Coupling’ here refers to the idea that the captured image “randomly biases” the cross-correlation.

So, bearing in mind this additional item of signal coupling, the identification process now estimates the signal amplitude of each and every individual embedded code signal (as opposed to taking the cross-correlation result of step 12, FIG. 3). If our identification signal exists in the suspect picture, the amplitudes thus found will split into a polarity with positive amplitudes being assigned a ‘1’ and negative amplitudes being assigned a

The new image is applied to the fast fourier transform routine and a scale factor is eventually found which minimizes the integrated high frequency content of the new image. This general type of search operation with its minimization of a particular quantity is exceedingly common. The scale factor thus found is the sought-for “amplitude.” Refinements which are contemplated but not yet implemented are where the coupling of the higher derivatives of the acquired image and the embedded codes are estimated and removed from the calculated scale factor. In other words, certain bias effects from the coupling mentioned earlier are present and should be eventually accounted for and removed both through theoretical and empirical experimentation.

Use and Improvements in the Detection of Signal or Image Alteration

Apart from the basic need of identifying a signal or image as a whole, there is also a rather ubiquitous need to detect possible alterations to a signal or image. The following section describes how the foregoing embodiment, with certain modifications and improvements, can be used as a powerful tool in this area. The potential scenarios and applications of detecting alterations are innumerable.

To first summarize, assume that we have a given signal or image which has been positively identified using the basic methods outlined above. In other words, we know its N-bit identification word, its individual embedded code signals, and its composite embedded code. We can then fairly simply create a spatial map of the composite code's amplitude within our given signal or image. Furthermore, we can divide this amplitude map by the known composite code's spatial amplitude, giving a normalized map, i.e. a map which should fluctuate about some global mean value. By simple examination of this map, we can visually detect any areas which have been significantly altered wherein the value of the normalized amplitude dips below some statistically set threshold based purely on typical noise and corruption (error).

The details of implementing the creation of the amplitude map have a variety of choices. One is to perform the same procedure which is used to determine the signal amplitude as described above, only now we step and repeat the multiplication of any given area of the signal/image with a gaussian weight function centered about the area we are investigating.

Universal Versus Custom Codes

The disclosure thus far has outlined how each and every source signal has its own unique set of individual embedded code signals. This entails the storage of a significant amount of additional code information above and beyond the original, and many applications may merit some form of economizing.

One such approach to economizing is to have a given set of individual embedded code signals be common to a batch of source materials. For example, one thousand images can all utilize the same basic set of individual embedded code signals. The storage requirements of these codes then become a small fraction of the overall storage requirements of the source material.

Furthermore, some applications can utilize a universal set of individual embedded code signals, i.e., codes which remain the same for all instances of distributed material. This type of requirement would be seen by systems which wish to hide the N-bit identification word itself, yet have standardized equipment be able to read that word. This can be used in systems which make go/no go decisions at point-of-read locations. The potential drawback to this set-up is that the universal codes are more prone to be sleuthed or stolen; therefore they will not be as secure as the apparatus and methodology of the previously disclosed arrangement. Perhaps this is just the difference between ‘high security’ and ‘air-tight security,’ a distinction carrying little weight with the bulk of potential applications.

Use in Printing, Paper, Documents, Plastic Coated Identification Cards, and Other Material Where Global Embedded Codes Can Be Imprinted

The term ‘signal’ is often used narrowly to refer to digital data information, audio signals, images, etc. A broader interpretation of ‘signal,’ and the one more generally intended, includes any form of modulation of any material whatsoever. Thus, the micro-topology of a piece of common paper becomes a ‘signal’ (e.g. it height as a function of x-y coordinates). The reflective properties of a flat piece of plastic (as a function of space also) becomes a signal. The point is that photographic emulsions, audio signals, and digitized information are not the only types of signals capable of utilizing the principles described herein.

As a case in point, a machine very much resembling a braille printing machine can be designed so as to imprint unique ‘noise-like’ indentations as outlined above. These indentations can be applied with a pressure which is much smaller than is typically applied in creating braille, to the point where the patterns are not noticed by a normal user of the paper. But by following the steps of the present disclosure and applying them via the mechanism of micro-indentations, a unique identification code can be placed onto any given sheet of paper, be it intended for everyday stationary purposes, or be it for important documents, legal tender, or other secured material.

The reading of the identification material in such an embodiment generally proceeds by merely reading the document optically at a variety of angles. This would become an inexpensive method for deducing the micro-topology of the paper surface. Certainly other forms of reading the topology of the paper are possible as well.

In the case of plastic encased material such as identification cards, e.g. driver's licenses, a similar braille-like impressions machine can be utilized to imprint unique identification codes. Subtle layers of photoreactive materials can also be embedded inside the plastic and ‘exposed.’

It is clear that wherever a material exists which is capable of being modulated by ‘noise-like’ signals, that material is an appropriate carrier for unique identification codes and utilization of the principles disclosed herein. All that remains is the matter of economically applying the identification information and maintaining the signal level below an acceptability threshold which each and every application will define for itself.

Real Time Encoder

While the first class of embodiments most commonly employs a standard microprocessor or computer to perform the encodation of an image or signal, it is possible to utilize a custom encodation device which may be faster than a typical Von Neuman-type processor. Such a system can be utilized with all manner of serial data streams.

Music and videotape recordings are examples of serial data streams—data streams which are often pirated. It would assist enforcement efforts if authorized recordings were encoded with identification data so that pirated knock-offs could be traced to the original from which they were made.

Piracy is but one concern driving the need for applicant's technology. Another is authentication. Often it is important to confirm that a given set of data is really what it is purported to be (often several years after its generation).

To address these and other needs, the system 200 of FIG. 5 can be employed. System 200 can be thought of as an identification coding black box 202. The system 200 receives an input signal (sometimes termed the “master” or “unencoded” signal) and a code word, and produces (generally in real time) an identification-coded output signal. (Usually, the system provides key data for use in later decoding.)

The contents of the “black box” 202 can take various forms. An exemplary black box system is shown in FIG. 6 and includes a look-up table 204, a digital noise source 206, first and second scalers 208, 210, an adder/subtracter 212, a memory 214, and a register 216.

The input signal (which in the illustrated embodiment is an 8-20 bit data signal provided at a rate of one million samples per second, but which in other embodiments could be an analog signal if appropriate A/D and D/A conversion is provided) is applied from an input 218 to the address input 220 of the look-up table 204. For each input sample (i.e. look-up table address), the table provides a corresponding 8-bit digital output word. This output word is used as a scaling factor that is applied to one input of the first scaler 208.

The first scaler 208 has a second input, to which is applied an 8-bit digital noise signal from source 206. (In the illustrated embodiment, the noise source 206 comprises an analog noise source 222 and an analog-to-digital converter 224 although, again, other implementations can be used.) The noise source in the illustrated embodiment has a zero mean output value, with a full width half maximum (FWHM) of 50-100 digital numbers (e.g. from −75 to +75).

The first scaler 208 multiplies the two 8-bit words at its inputs (scale factor and noise) to produce—for each sample of the system input signal—a 16-bit output word. Since the noise signal has a zero mean value, the output of the first scaler likewise has a zero mean value.

The output of the first scaler 208 is applied to the input of the second scaler 210. The second scaler serves a global scaling function, establishing the absolute magnitude of the identification signal that will ultimately be embedded into the input data signal. The scaling factor is set through a scale control device 226 (which may take a number of forms, from a simple rheostat to a graphically implemented control in a graphical user interface), permitting this factor to be changed in accordance with the requirements of different applications. The second scaler 210 provides on its output line 228 a scaled noise signal. Each sample of this scaled noise signal is successively stored in the memory 214.

(In the illustrated embodiment, the output from the first scaler 208 may range between −1500 and +1500 (decimal) while the output from the second scaler 210 is in the low single digits, (such as between −2 and +2).)

Register 216 stores a multi-bit identification code word. In the illustrated embodiment this code word consists of 8 bits, although larger code words (up to hundreds of bits) are commonly used. These bits are referenced, one at a time, to control how the input signal is modulated with the scaled noise signal.

In particular, a pointer 230 is cycled sequentially through the bit positions of the code word in register 216 to provide a control bit of “0” or “1” to a control input 232 of the adder/subtracter 212. If, for a particular input signal sample, the control bit is a “1”, the scaled noise signal sample on line 10
where =\=is not equal to. This is somewhat an abstract notion to introduce at this point in the disclosure and will become more clear as FIG. 15 is discussed. The general idea, however, is that there will be a variety of algebras that can be used to optimize the pass-through of “invisible” signatures through compression procedures. Clearly, the same principles as depicted in FIG. 15 also work on still images and the JPEG or any other still image compression standard.

Turning now to the details of FIG. 15, we begin with the simple stepping through of all Z frames of a movie or video. For a two hour movie played at 30 frames per second, Z turns out to be (30*2*60*60) or 216,000. The inner loop of 700, 702 and 704 merely mimics FIG. 13's steps. The logo frame optionally can change during the stepping through frames. The two arrows emanating from the box 704 represent both the continuation of the loop 750 and the depositing of output frames into the rendered master Snowy Image 752.

To take a brief but potentially appropriate digression at this point, the use of the concept of a Markov process brings certain clarity to the discussion of optimizing the engineering implementation of the methods of FIG. 15. Briefly, a Markov process is one in which a sequence of events takes place and in general there is no memory between one step in the sequence and the next. In the context of FIG. 15 and a sequence of images, a Markovian sequence of images would be one in which there is no apparent or appreciable correlation between a given frame and the next. Imagine taking the set of all movies ever produced, stepping one frame at a time and selecting a random frame from a random movie to be inserted into an output movie, and then stepping through, say, one minute or 1800 of these frames. The resulting “movie” would be a fine example of a Markovian movie. One point of this discussion is that depending on how the logo frames are rendered and depending on how the encryption/scrambling step 702 is performed, the Master Snowy Movie 752 will exhibit some generally quantifiable degree of Markovian characteristics. The point of this point is that the compression procedure itself will be affected by this degree of Markovian nature and thus needs to be accounted for in designing the process of FIG. 15. Likewise, and only in general, even if a fully Markovian movie is created in the High Brightness Master Snowy Movie, 752, then the processing of compressing and decompressing that movie 752, represented as the MPEG box 754, will break down some of the Markovian nature of 752 and create at least a marginally non-Markovian compressed master Snowy Movie 756. This point will be utilized when the disclosure briefly discusses the idea of using multiple frames of a video stream in order to find a single N-bit identification word, that is, the same N-bit identification word may be embedded into several frames of a movie, and it is quite reasonable to use the information derived from those multiple frames to find that single N-bit identification word. The non-Markovian nature of 756 thus adds certain tools to reading and recognizing the invisible signatures. Enough to this tangent.

With the intent of preconditioning the ultimately utilized Master Snowy Movie 756, we now send the rendered High Brightness Master Snowy Movie 752 through both the MPEG compression AND decompression procedure 754. With the caveat previously discussed where it is acknowledged that the MPEG compression process is generally not distributive, the idea of the step 754 is to crudely segregate the initially rendered Snowy Movie 752 into two components, the component which survives the compression process 754 which is 756, and the component which does not survive, also crudely estimated using the difference operation 758 to produce the “Cheap Master Snowy Movie” 760. The reason use is made of the deliberately loose term “Cheap” is that we can later add this signature signal as well to a distributable movie, knowing that it probably won't survive common compression processes but that nevertheless it can provide “cheap” extra signature signal energy for applications or situations which will never experience compression. [Thus it is at least noted in FIG. 15]. Back to FIG. 15 proper, we now have a rough cut at signatures which we know have a higher likelihood of surviving intact through the compression process, and we use this “Compressed Master Snowy Movie” 756 to then go through this procedure of being scaled down 764, added to the original movie 766, producing a candidate distributable movie 770, then compared to the original movie (768) to ensure that it meets whatever commercially viable criteria which have been set up (i.e. the acceptable perceived noise level). The arrow from the side-by-side step 768 back to the scale down step 764 corresponds quite directly to the “experiment visually . . . ” step of FIG. 2, and the gain control 226 of FIG. 6. Those practiced in the art of image and audio information theory can recognize that the whole of FIG. 15 can be summarized as attempting to pre-condition the invisible signature signals in such a way that they are better able to withstand even quite appreciable compression. To reiterate a previously mentioned item as well, this idea equally applies to ANY such pre-identifiable process to which an image, and image sequence, or audio track might be subjected. This clearly includes the JPEG process on still images.

Additional Elements of the Realtime Encoder Circuitry

It should be noted that the method steps represented in FIG. 15, generally following from box 750 up through the creation of the compressed master snowy movie 756, could with certain modification be implemented in hardware. In particular, the overall analog noise source 206 in FIG. 6 could be replaced by such a hardware circuit. Likewise the steps and associated procedures depicted in FIG. 13 could be implemented in hardware and replace the analog noise source 206.

Recognition based on more than one frame: non-Markovian signatures

As noted in the digression on Markov and non-Markov sequences of images, it is pointed out once again that in such circumstances where the embedded invisible signature signals are non-Markovian in nature, i.e., that there is some correlation between the master snowy image of one frame to that of the next, AND furthermore that a single N-bit identification word is used across a range of frames and that the sequence of N-bit identification words associated with the sequence of frames is not Markovian in nature, then it is possible to utilize the data from several frames of a movie or video in order to recognize a single N-bit identification word. All of this is a fancy way of saying that the process of recognizing the invisible signatures should use as much information as is available, in this case translating to multiple frames of a motion image sequence.

Header Verification

The concept of the “header” on a digital image or audio file is a well established practice in the art. The top of FIG. 16 has a simplified look at the concept of the header, wherein a data file begins with generally a comprehensive set of information about the file as a whole, often including information about who the author or copyright holder of the data is, if there is a copyright holder at all. This header 800 is then typically followed by the data itself 802, such as an audio stream, DIAAppendix B

A software version of a steganographic marking/decoding “plug-in” for use with Adobe Photoshop software, presented as commented source code, is included in the file of this patent on a compact disc in a file named Appendix B.txt, which is incorporated by reference. The code was written for compilation with Microsoft's Visual C++ compiler, version 4.0, and can be understood by those skilled in the art.

If marking of images becomes widespread (e.g. by software compatible with Adobe's image processing software), a user of such software can decode the embedded data from an image and consult a public registry to identify the proprietor of the image. In some embodiments, the registry can serve as the conduit through which appropriate royalty payments are forwarded to the proprietor for the user's use of an image. (In an illustrative embodiment, the registry is a server on the Internet, accessible via the World Wide Web, coupled to a database. The database includes detailed information on catalogued images (e.g. name, address, phone number of proprietor, and a schedule of charges for different types of uses to which the image may be put), indexed by identification codes with which the images themselves are encoded. A person who decodes an image queries the registry with the codes thereby gleaned to obtain the desired data and, if appropriate, to forward electronic payment of a copyright royalty to the image's proprietor.)

Particular Data Formats

While the foregoing steganography techniques are broadly applicable, their commercial acceptance will be aided by establishment of standards setting forth which pixels/bit cells represent what. The following discussion proposes one set of possible standards. For expository convenience, this discussion focuses on decoding of the data; encoding follows in a straightforward manner.

Referring to FIG. 42, an image 1202 includes a plurality of tiled “signature blocks” 1204. (Partial signature blocks may be present at the image edges.) Each signature block 1204 includes an 8×8 array of sub-blocks 1206. Each sub-block 1206 includes an 8×8 array of bit cells 1208. Each bit cell comprises a 2×2 array of “bumps” 1210. Each bump 1210, in turn, comprises a square grouping of 16 individual pixels 1212.

The individual pixels 1212 are the smallest quanta of image data. In this arrangement, however, pixel values are not, individually, the data carrying elements. Instead, this role is served by bit cells 1208 (i.e. 2×2 arrays of bumps 1210). In particular, the bumps comprising the bits cells are encoded to assume one of the two patterns shown in FIG. 41. As noted earlier, the pattern shown in FIG. 41A represents a “0” bit, while the pattern shown in FIG. 41B represents a “1” bit. Each bit cell 1208 (64 image pixels) thus represents a single bit of the embedded data. Each sub-block 1206 includes 64 bit cells, and thus conveys 64 bits of embedded data.

(The nature of the image changes effected by the encoding follows the techniques set forth above under the heading MORE ON PERCEPTUALLY ADAPTIVE SIGNING; that discussion is not repeated here.)

In the illustrated embodiment, the embedded data includes two parts: control bits and message bits. The 16 bit cells 1208A in the center of each sub-block 1206 serve to convey 16 control bits. The surrounding 48 bit cells 1208B serve to convey 48 message bits. This 64-bit chunk of data is encoded in each of the sub-blocks 1206, and is repeated 64 times in each signature block 1204.

A digression: in addition to encoding of the image to redundantly embed the 64 control/message bits therein, the values of individual pixels are additionally adjusted to effect encoding of subliminal graticules through the image. In this embodiment, the graticules discussed in conjunction with FIG. 29A are used, resulting in an imperceptible texturing of the image. When the image is to be decoded, the image is transformed into the spatial domain, the Fourier-Mellin technique is applied to match the graticule energy points with their expected positions, and the processed data is then inverse-transformed, providing a registered image ready for decoding. (The sequence of first tweaking the image to effect encoding of the subliminal graticules, or first tweaking the image to effect encoding of the embedded data, is not believed to be critical. As presently practiced, the local gain factors (discussed above) are computed; then the data is encoded; then the subliminal graticule encoding is performed. (Both of these encoding steps make use of the local gain factors.))

Returning to the data format, once the encoded image has been thus registered, the locations of the control bits in sub-block 1206 are known. The image is then analyzed, in the aggregate (i.e. considering the “northwestern-most” sub-block 1206 from each signature block 1204), to determine the value of control bit #1 (represented in sub-block 1206 by bit cell 1208Aa). If this value is determined (e.g. by statistical techniques of the sort detailed above) to be a “1,” this indicates that the format of the embedded data conforms to the standard detailed herein (the Digimarc Beta Data Format).

According to this standard, control bit #2 (represented by bit cells 1208Ab) is a flag indicating whether the image is copyrighted. Control bit #3 (represented by bit cells 1208Ac) is a flag indicating whether the image is unsuitable for viewing by children. Certain of the remaining bits are used for error detection/correction purposes.

The 48 message bits of each sub block 1206 can be put to any use; they are not specified in this format. One possible use is to define a numeric “owner” field and a numeric “image/item” field (e.g. 24 bits each).

If this data format is used, each sub-block 1206 contains the entire control/message data, so same is repeated 64 times within each signature block of the image. If control bit #1 is not a “1,” then the format of the embedded data does not conform to the above described standard. In this case, the reading software analyzes the image data to determine the value of control bit #4. If this bit is set (i.e. equal to “1”), this signifies an embedded ASCII message.

The reading software then examines control bits #5 and #6 to determine the length of the embedded ASCII message.

If control bits #5 and #6 both are “0,” this indicates the ASCII message is 6 characters in length. In this case, the 48 bit cells 1208B surrounding the control bits 1208A are interpreted as six ASCII characters (8 bits each). Again, each sub-block 1206 contains the entire control/message data, so same is repeated 64 times within each signature block 1204 of the image.

If control bit #5 is “0” and control bit #6 is “1,” this indicates the embedded ASCII message is 14 characters in length. In this case, the 48 bit cells 1208B surrounding the control bits 1208A are interpreted as the first six ASCII characters. The 64 bit cells 1208 of the immediately-adjoining subblock 1220 are interpreted as the final eight ASCII characters.

Note that in this arrangement, the bit-cells 1208 in the center of sub-block 1220 are not interpreted as control bits. Instead, the entire sub-block serves to convey additional message bits. In this case there is just one group of control bits for two sub-blocks.

Also note than in this arrangement, pairs of sub-blocks 1206 contains the entire control/message data, so same is repeated 32 times within each signature block 1204 of the image.

Likewise if control bit #5 is “1” and control bit #6 is “0.” This indicates the embedded ASCII message is 30 characters in length. In this case, 2×2 arrays of sub-blocks are used for each representation of the data. The 48 bit cells 1208B surrounding control bits 1208A are interpreted as the first six ASCII characters. The 64 bit cells of each of adjoining block 1220 are interpreted as representing the next 8 additional characters. The 64 bits cells of sub-block 1222 are interpreted as representing the next 8 characters. And the 64 bit cells of sub-block 1224 are interpreted as representing the final 8 characters. In this case, there is just one group of control bits for four sub-blocks. And the control/message data is repeated 16 times within each signature block 1204 of the image.

If control bits #5 and #6 are both “1”s, this indicates an ASCII message of programmable length. In this case, the reading software examines the first 16 bit cells 1208B surrounding the control bits. Instead of interpreting these bit cells as message bits, they are interpreted as additional control bits (the opposite of the case described above, where bit cells normally used to represent control bits represented message bits instead). In particular, the reading software interprets these 16 bits as representing, in binary, the length of the ASCII message. An algorithm is then applied to this data (matching a similar algorithm used during the encoding process) to establish a corresponding tiling pattern (i.e. to specify which sub-blocks convey which bits of the ASCII message, and which convey control bits.)

In this programmable-length ASCII message case, control bits are desirably repeated several times within a single representation of the message so that, e.g., there is one set of control bits for approximately every 24 ASCII characters. To increase packing efficiency, the tiling algorithm can allocate (divide) a sub-block so that some of its bit-cells are used for a first representation of the message, and others are used for another representation of the messages.

Reference was earlier made to beginning the decoding of the registered image by considering the “northwestern-most” sub-block 1206 in each signature block 1204. This bears elaboration.

Depending on the data format used, some of the sub-blocks 1206 in each signature block 1204 may not include control bits. Accordingly, the decoding software desirably determines the data format by first examining the “northwestern-most” sub-block 1206 in each signature block 1204; the 16 bits cells in the centers of these sub-blocks will reliably represent control bits. Based on the value(s) of one or more of these bits (e.g. the Digimarc Beta Data Format bit), the decoding software can identify all other locations throughout each signature block 1204 where the control bits are also encoded (e.g. at the center of each of the 64 sub-blocks 1206 comprising a signature block 1204), and can use the larger statistical base of data thereby provided to extract the remaining control bits from the image (and to confirm, if desired, the earlier control bit(s) determination). After all control bits have thereby been discerned, the decoding software determines (from the control bits) the mapping of message bits to bit cells throughout the image.

To reduce the likelihood of visual artifacts, the numbering of bit cells within sub-blocks is alternated in a checkerboard-like fashion. That is, the “northwestern-most” bit cell in the “northwestern-most” sub-block is numbered “0.” Numbering increases left to right, and successively through the rows, up to bit cell 63. Each sub-block diametrically adjoining one of its corners (i.e. sub-block 1224) has the same ordering of bit cells. But sub-blocks adjoining its edges (i.e. sub-blocks 1220 and 1222) have the opposite numbering. That is, the “northwestern-most” bit cell in sub-blocks 1220 and 1222 is numbered “63.” Numbering decreases left to right, and successively through the rows, down to 0. Likewise throughout each signature block 1204.

In a variant of the Digimarc beta format, a pair of sub-blocks is used for each representation of the data, providing 128 bit cells. The center 16 bit cells 1208 in the first sub-block 1206 are used to represent control bits. The 48 remaining bit cells in that sub-block, together with all 64 bit cells 1208 in the adjoining sub-block 1220, are used to provide a 112-bit message field. Likewise for every pair of sub-blocks throughout each signature block 1204. In such an arrangement, each signature block 1204 thus includes 32 complete representations of the encoded data (as opposed to 64 representations in the earlier-described standard). This additional length allows encoding of longer data strings, such as a numeric IP address (e.g. URL).

Obviously, numerous alternative data formats can be designed. The particular format used can be indicated to the decoding software by values of one or more control bits in the encoded image.

In the Appendix B software, the program SIGN_PUBLIC.CPP effects encoding of an image using a signature block/sub-block/bit cell arrangement like that detailed above. As of this writing, the corresponding decoding software has not yet been written, but its operation is straight-forward given the foregoing discussion and the details in the SIGN-PUBLIC.CPP software.

Other Applications

Before concluding, it may be instructive to review some of the other fields where principles of applicant's technology can be employed.

One is smart business cards, wherein a business card is provided with a photograph having unobtrusive, machine-readable contact data embedded therein. (The same function can be achieved by changing the surface microtopology of the card to embed the data therein.)

Another promising application is in content regulation. Television signals, images on the internet, and other content sources (audio, image, video, etc.) can have data indicating their “appropriateness” (i.e. their rating for sex, violence, suitability for children, etc.) actually embedded in the content itself rather than externally associated therewith. Television receivers, internet surfing software, etc., can discern such appropriateness ratings (e.g. by use of universal code decoding) and can take appropriate action (e.g. not permitting viewing of an image or video, or play-back of an audio source).

In a simple embodiment of the foregoing, the embedded data can have one or more “flag” bits, as discussed earlier. One flag bit can signify “inappropriate for children.” (Others can be, e.g., “this image is copyrighted” or “this image is in the public domain.”) Such flag bits can be in a field of control bits distinct from an embedded message, or can—themselves—be the message. By examining the state of these flag bits, the decoder software can quickly apprise the user of various attributes of the image.

(As discussed earlier, control bits can be encoded in known locations in the image—known relative to the subliminal graticules—and can indicate the format of the embedded data (e.g. its length, its type, etc.) As such, these control bits are analogous to data sometimes conveyed in prior art file headers, but in this case they are embedded within an image, instead of prepended to a file.)

The field of merchandise marking is generally well served by familiar bar codes and universal product codes. However, in certain applications, such bar codes are undesirable (e.g. for aesthetic considerations, or where security is a concern). In such applications, applicant's technology may be used to mark merchandise, either through an innocuous carrier (e.g. a photograph associated with the product), or by encoding the microtopology of the merchandise's surface, or a label thereon.

There are applications—too numerous to detail—in which steganography can advantageously be combined with encryption and/or digital signature technology to provide enhanced security.

Medical records appear to be an area in which authentication is important. Steganographic principles—applied either to film-based records or to the microtopology of documents—can be employed to provide some protection against tampering.

Many industries, e.g. automobile and airline, rely on tags to mark critical parts. Such tags, however, are easily removed, and cap often be counterfeited. In applications wherein better security is desired, industrial parts can be steganographically marked to provide an inconspicuous identification/authentication tag.

In various of the applications reviewed in this specification, different messages can be steganographically conveyed by different regions of an image (e.g. different regions of an image can provide different internet URLs, or different regions of a photocollage can identify different photographers). Likewise with other media (e.g. sound).

Some software visionaries look to the day when data blobs will roam the datawaves and interact with other data blobs. In such an era, it will be necessary for such blobs to have robust and incorruptible ways of identifying themselves. Steganographic techniques again bold much promise here.

Finally, message changing codes—recursive systems in which steganographically encoded messages actually change underlying steganographic code patterns—offer new levels of sophistication and security. Such message changing codes are particularly well suited to applications such as plastic cash cards where time-changing elements are important to enhance security.

Again while applicant prefers the particular forms of steganographic encoding detailed above, the diverse applications disclosed in this specification can largely be practiced with other steganographic marking techniques, many of which are known in the prior art. Likewise, while the specification has focused on applications of this technology to images, the principles thereof are generally equally applicable to embedding such information in audio, physical media, or any other carrier of information.

Finally, while the specification has been illustrated with particular embodiments, it will be recognized that elements, components and steps from these embodiments can be recombined in different arrangements to serve different needs and applications, as will be readily apparent to those of ordinary skill in the art.

In view of the wide variety of implementations and applications to which the principles of this technology can be put, it should be apparent that the detailed embodiments are illustrative only and in no way limit the scope of my invention. Instead, I claim as my invention all such embodiments as come within the scope and spirit of the following claims and equivalents thereto.

Claims

1. A method for surveying distribution of proprietary empirical data sets, such as audio, image, or video data, on computer sites accessible via the internet, comprising:

automatically downloading data, including empirical data sets, from a plurality of computer sites over the internet;

for each of a plurality of empirical data sets obtained by said downloading operation, automatically screening same to identify the potential presence of identification data steganographically encoded therein;

for each of a plurality of empirical data sets screened by said screening operation, discerning identification data, if any, steganographically encoded therein; and

generating a report identifying steganographically encoded empirical data sets identified by the foregoing steps, and the site from which each was downloaded;

wherein there is calibration data steganographically encoded within at least one empirical data set, said calibration data having one or more known properties facilitating identification thereof during the discerning step;

the method including identifying the calibration data within the empirical data set and using data obtained thereby to aid in discerning the identification data from the empirical data set;

wherein the empirical data set has been corrupted since being encoded, said corruption including a process selected from the group consisting of: misregistration and scaling of the empirical data set;

the method further including using said data to compensate for said corruption, wherein the identification data can nonetheless be recovered from the empirical data set notwithstanding said corruption.

2. The method of claim 1 which includes providing a master code signal, and using said code signal is discerning said steganographically encoded identification data from said screened empirical data sets.

3. The method of claim 2 in which said master code signal has the appearance of unpatterned snow if represented in the a pixel domain.

4. The method of claim 1 in which said discerning of identification data from said downloaded empirical data is accomplished without previous knowledge of the audio, image, or video information represented thereby.

5. The method of claim 1 which includes identifying proprietors of empirical data sets by reference to identification data steganographically discerned therefrom, and reporting to said proprietors the sites from which their empirical data sets were downloaded.

6. The method of claim 5 in which said identification data includes information in addition to data identifying said proprietor, and the method includes providing said additional data to said proprietors.

7. The method of claim 5 in which said identification data is a serial number index into a registry database containing names and contact information for proprietors identified by said identification data.

8. The method of claim 1 in which the empirical data sets include image data, and the method includes:

converting said image data to pixel form, if not already in said form; and

performing a plurality of statistical analyses on said pixel form image data to discern the identification data therefrom.

9. The method of claim 8 in which each statistical analysis includes analyzing a collection of spaced apart pixels to decode a single, first bit of the identification data therefrom, said analysis to decode the first bit encompassing not just said spaced apart pixels, but also pixels adjacent thereto, said adjacent pixels not being encoded with said first bit.

10. A method for surveying distribution of proprietary empirical data sets on computer sites accessible via the internet, comprising:

providing a master code signal useful for detecting steganographic coding within empirical data sets;

automatically downloading data, including empirical data sets, from a plurality of computer sites over the internet;

for each of a plurality of empirical data sets obtained by said downloading operation, discerning certain identification data, if any, steganographically encoded therein, said discerning employing said master code signal as a decoding key; and

generating a report identifying steganographically encoded empirical data sets identified by the foregoing steps, and the site from which each was downloaded;

wherein there is calibration data steganographically encoded within at least one empirical data set, said calibration data having one or more known properties facilitating identification thereof during the discerning step;

the method including identifying the calibration data within the empirical data set and using data obtained thereby to aid in discerning the identification data from the empirical data set;

wherein the empirical data set has been corrupted since being encoded, said corruption including a process selected from the group consisting of: misregistration and scaling of the empirical data set;

the method further including using said data to compensate for said corruption, wherein the identification data can nonetheless be recovered from the empirical data set notwithstanding said corruption.

11. The method of claim 10 which includes automatically screening each of a plurality of said empirical data sets obtained by said downloading operation, to identify the potential presence of identification data steganographically encoded therein and, for those data sets that pass said screening process, discerning identification data, if any, steganographically encoded therein.

12. The method of claim 10 in which said master code signal has the appearance of unpatterned snow if represented in the a pixel domain.

13. The method of claim 10 in which said discerning of identification data from said downloaded empirical data is accomplished without previous knowledge of the audio, image, or video information represented thereby.

14. The method of claim 10 which includes identifying proprietors of empirical data sets by reference to identification data steganographically discerned therefrom, and reporting to said proprietors the sites from which their empirical data sets were downloaded.

15. The method of claim 14 in which said identification data includes information in addition to data identifying said proprietor, and the method includes providing said additional data to said proprietors.

16. The method of claim 14 in which said identification data is a serial number index into a registry database containing names and contact information for proprietors identified by said identification data.

17. The method of claim 10 in which the empirical data sets include image data, and the method includes:

converting said image data to pixel form, if not already in said form; and

performing a plurality of statistical analyses on said pixel form image data to discern the identification data therefrom.

18. The method of claim 17 in which each statistical analysis includes analyzing a collection of spaced apart pixels to decode a single, first bit of the identification data therefrom, said analysis to decode the first bit encompassing not just said spaced apart pixels, but also pixels adjacent thereto, said adjacent pixels not being encoded with said first bit.