AU6936398A

AU6936398A - Method and apparatus for camouflaging data

Info

Publication number: AU6936398A
Application number: AU69363/98A
Authority: AU
Inventors: Anthony Tung Shuen Ho
Original assignee: Datamark Technologies Pte Ltd
Current assignee: Datamark Technologies Pte Ltd
Priority date: 1998-03-18
Filing date: 1998-03-18
Publication date: 1999-10-11
Also published as: EP1062585A1; JP2002507771A; CA2319601A1; WO1999048006A1

Description

WO 99/48006 1 PCT/SG98/00023 METHOD AND APPARATUS FOR CAMOUFLAGING DATA Field of the invention The present invention relates generally to encoding methods, and more 5 particularly to a method and apparatus for steganographic encoding of data. The encoding method is espartially suited to digital camouflaging of audio, still image and video data and it will be convenient to describe the method and associated apparatus in relation to that example application. It should be appreciated, however, that the invention is intended for broader application and 10 use. For example, the encoding method of the present invention may also be used in the field of data encryption. Background of the invention The tremendous growth in multimedia products and services provided via 15 the internet has led to the need for copyright authentication and for protecting data integrity. In the past few years, a number of digital watermarking techniques have been developed for the purpose of resolving legal use issues associated with copyright information on the internet. A number of digital watermarking techniques have recently been patented. 20 Examples of these include US Patent 5,636,292 to Rhoads (1997) and US Patent 5,659,726 to Sandford and Handel (1997). Rhoads discloses methods to impress an identification code on a carrier, such as an electronic data signal or a physical medium, in a manner that permits the identification code to be later discerned and the carrier thereby identified. Sandford and Handel disclose a method of 25 embedding auxiliary information into host data, such as a photograph, television signal, facsimile transmission, or identification card. The method operates by manipulating a noise component of the host data in accordance with the auxiliary information. Throughout this specification the word "steganography" is intended to 30 denote any of various methods seeking to conceal the existence of a message within some other medium such that any unintended party who intercepts the WO 99/48006 PCT/SG98/00023 2 medium does not know it contains the concealed message and does not therefore obtain the information contained in the message. Digital watermarking is one example of a steganographic method used to embed secondary data, such as text, in primary or host data, such as a digitized image. 5 The word "cryptography" is intended to denote any of various techniques seeking to hide information contained in a message such that when the message is transmitted from a sender to a receiver any third party intercepting the message cannot read it or extract the information contained in the message. In the present context, the word "camouflaging" is intended to denote a 10 steganographic encoding method which conceals secondary data by utilizing a combination of primary or host data and key data. The encoding method may not actually embed information in the primary data, as in watermarking, but may utilize information elements existing in the primary data. Many prior art digital watermarking techniques, including the techniques 15 disclosed in the above US patents, are only able to conceal limited information, such as a few logical bits (ie. "1" and "0") or a few characters (eg. "A12"), in the host data. However, to record detailed ownership information for a host work in which copyright subsists, such as an image of Singapore, an entire message or sentence may need to be concealed in, or associated with, the host data. For 20 example, the sentence "Digital image of Singapore is the property of Mr John Tan, dated 16 December 1997" may provide more conclusive proof as to true ownership of the host work than having to rely on just a simple code to assess copyright infringement. There therefore remains a need for a steganographic encoding method 25 and apparatus which may allow a relatively long string of secondary data (such as text) to be camouflaged in association with primary data (such as image, audio or video data) whilst producing minimal degradation of the primary or host data. In some embodiments, the primary data may be absolutely unaltered by the encoding process. 30 Besides the above mentioned application on the internet, many potential consumer, commercial and service applications may benefit from the use of WO 99/48006 PCT/SG98/00023 3 digital camouflaging/watermarking technology, both for copyright protection purposes and for secure transmission of information. These applications include encoding of text containing ownership identification or attribute information associated with digital still or video cameras, copyright protection and royalty 5 tracking of sound recordings in the music industry. Commercial and service sectors may also benefit from secure transmission and reception of sensitive information that could be camouflaged in normal data streams transmitted through an open channel. 10 Summary of the Invention In one aspect, the present invention provides a method of steganographic encoding including the steps of: (A) providing primary data containing a plurality of addressable first data elements; 15 (B) providing secondary data containing a plurality of second data elements; and (C) for each second data element: (C1) searching for a first data element which matches content of the second data element, and 20 (C2) generating a key element including the address of the matching first data element when a match is found for the content of the second data element. When a match is found for the content of the second data element, the address of the matching first data element may be stored as a key element 25 associated with the second data element. A string of key elements (containing addresses) may be associated with a corresponding string of second data elements. The string of key elements may specify where in the primary data each second data element of the secondary data is to be found. The secondary data may in this context be considered to be camouflaged in the primary data and the 30 key data. Unique key data which is generated for given primary and secondary data may be stored for use in a complementary decoding method which is WO 99/48006 PCT/SG98/00023 4 described below. The primary and secondary data may be represented digitally. However, in its broadest aspects, the invention is applicable to any form of data representation or notation using any convenient set of symbols. The primary data 5 may represent a still image, motion video, audio, text or other type of information. Likewise, the secondary data may represent a still image, motion video, audio, text or other information. In a preferred form of the invention, the secondary data includes a text message and each second data element includes an alphanumeric character. The alphanumeric characters may be used to compose 10 the text message. In a typical application of the invention the text message may include copyright information relating to the image, video, audio, etc. contained in the primary data. In one embodiment, the text message may include one or more of the following: a title, an artist, a copyright holder, a body to which royalties should 15 be paid, and general terms of publisher distribution. The primary data may include a still image. The first data elements may be arranged in a two-dimensional array wherein each first data element defines a characteristic associated with a still image element. Typically, the first data elements are obtained from a stream of data representing a digitised still image. 20 The image may be obtained from a still digital camera, a computer game or other software, or other source. It may be a greyscale or color image, for example, and may be stored in any known format, eg. BMP, GIF, TIFF, or JPEG. The primary data may alternatively or additionally include motion video. The first data elements may be arranged in a three-dimensional array wherein 25 each first data element defines a characteristic associated with a motion video element. Typically, the first data elements are obtained from a stream of data representing digitised motion video. The digitised video may be obtained from a Video Compact Disc (VCD) player, a Laser Disc (LD) player, a computer game or other software, a Digital Versatile Disc (DVD) player or other source, and may be 30 stored in any known format, eg. MPEG. The primary data may alternatively or additionally include audio WO 99/48006 PCT/SG98/00023 5 information. The first data elements may be arranged in a one-dimensional array wherein each first data element defines a characteristic associated with a digital audio sample. Typically, the digital audio samples are obtained from a stream of data representing digitised sound or music. The digitised sound may be obtained 5 from a Compact Disc (CD) player, Digital Audio Tape (DAT) player, Laser Disc player, Video Compact Disc (VCD) player or other source, and may be stored in any known format eg. WAV, AIFF, etc.. In one embodiment, the digital audio samples are obtained from two streams of data representing two channels of digitised sound for stereo reproduction. 10 In the encoding method described above, the first and second data elements may be represented by integer values and step (C) may further include: when a match is not found for the content of a second data element; (C3) producing an adjusted second data element by incrementing or decrementing the integer value of the second data element, 15 (C4) searching for a first data element which matches the integer value of the adjusted second data element, and (C5) generating a key element including the address of the matching first data element when a match is found for the adjusted second data element, and replacing the content of the matching first data element 20 with the integer value of the second data element prior to producing the adjusted second data element. When a match is not found for the adjusted second data element the method may further include: (C6) producing a new adjusted second data element by 25 incrementing or decrementing the adjusted second data element and repeating steps (C4) and (C5) for the new adjusted second data element. For example, if a desired integer value, say, 105 is not found in the primary data, a search is conducted to locate values 104 or 106. If one of these values is found, the address of the first data element containing that value is stored in a 30 key element, and the content of the first data element (104 or 106) is replaced with the value 105. If on the other hand the values 104 or 106 are not found, a WO 99/48006 PCT/SG98/00023 6 search is conducted for values 103 or 107. This modification of the primary data may be considered to be a form of embedding of the secondary data in the primary data, in the watermarking sense, but is only performed in the relatively rare event that a desired integer value is not found in the primary data. It has 5 been found that these rarely occurring changes are imperceptible in the modified primary image, video or audio data. The modified version of the primary data may subsequently be stored for distribution via the internet or other means. The modified primary data contains the secondary data which can be extracted at any future time using the 10 associated key data generated during the encoding process. The encoding method may include "data shifting" steps prior to step (C). The data shifting steps may be required whenever the range for the secondary data does not fall within the range for the primary data. This may occur for example when the primary data is a digitised audio signal and the secondary data 15 is a text message. Similar data shifting steps may be applied to other types of data. The data shifting steps may include: determining a range for the contents of the first data elements, determining a range for the contents of the second data elements, comparing the range for the first data elements with the range for the 20 second data elements, shifting the contents of the second data elements when the range for the second data elements falls outside of the range for the first data elements, such that the range for the second data elements falls substantially within the range for the first data elements, and 25 using the shifted second data elements as the second data elements in step (C). Preferably, the step of determining a range for the contents of the first data elements includes: calculating a mean and standard deviation for the first data elements; and determining a lower limit for the first data elements based on the 30 mean and standard deviation. The step of determining a range for the contents of the second data elements may include establishing as a reference a minimum WO 99/48006 PCT/SG98/00023 7 value which can be attributed to the range of possible second data elements. The step of comparing may include calculating an offset value by subtracting the reference value from the lower limit, and the step of shifting may include adding the offset value to the contents of each second data element. Preferably the 5 offset value is stored with the key elements in the key data. It will be appreciated that when a data shifting offset is applied to the secondary data by the encoding method, the complementary decoding method should also include steps to reverse the data shifting offset. The offset value may be retrieved from the key data. 10 Compared to existing steganographic or watermarking techniques the present invention has the distinct advantage that long sentences of text may be camouflaged. Even with long text strings, the integrity of the primary data is in most cases not affected or compromised. The primary data may remain absolutely unchanged by the encoding method because the method utilises data 15 already present in the primary data. Even in those cases where specific values contained in the secondary data cannot be found in the primary data, and the primary data is modified to insert those values, the primary data may remain substantially unaltered. In another aspect, the present invention provides a method of 20 steganographic decoding of secondary data including a plurality of second data elements, said secondary data being encoded in key elements in association with primary data, said method including the steps of: (A) providing said primary data containing a plurality of addressable first data elements; 25 (B) providing said key elements, each key element including an address of a first data element; and (C) for each key element, generating a said second data element by extracting the content of the addressed first data element. In a further aspect, the present invention provides an apparatus for 30 steganographic encoding including: (A) means for providing primary data containing a plurality of WO 99/48006 PCT/SG98/00023 8 addressable first data elements; (B) means for providing secondary data containing a plurality of second data elements; (C) means for searching, for each second data element, a first data 5 element which matches content of the second data element, and (D) means for generating a key element including the address of the matching first data element when a match is found for the content of the second data element. In a still further aspect, the present invention provides an apparatus for 10 steganographic decoding of secondary data including a plurality of second data elements, said secondary data being encoded in key elements in association with primary data, said apparatus including: (A) means for providing said primary data containing a plurality of addressable first data elements; 15 (B) means for providing said key elements, each key element including an address of a first data element; and (C) means for generating a second data element for each key element by extracting the content of the addressed first data element. In a still further aspect of the present invention the method of encoding 20 described above may be applied to cryptography. In applications involving cryptographic encoding a hidden message may be transmitted from a sender to a receiver. The message may be encoded in the key elements which are generated by the method in association with the primary data. The message to be hidden in this case corresponds to the second data elements. The hidden 25 message may be decoded by the receiver from the key elements by utilizing the primary data in a complementary decoding method. Brief Description of the Drawings The accompanying drawings, which are incorporated into and constitute 30 part of the description of the invention, illustrate embodiments of the invention and serve to explain the principles thereof. It is to be understood, however, that WO 99/48006 PCT/SG98/00023 9 the drawings and following detailed description are given for the purposes of illustration only and are not intended as a definition of the limits of the invention. In the drawings: Figure 1A illustrates a one dimensional data array containing integer 5 values representing a digitised audio signal; Figure 1B illustrates a two dimensional data array containing integer values representing a digital still image; Figure 2 illustrates an example of an audio signal; Figure 3 illustrates an example of a digital image; 10 Figure 4 illustrates a table of the ASCII character set; Figure 5 illustrates a typical greyscale histogram for a digital image; Figures 6A and 6B illustrate a two dimensional data array of 4 x 8 data elements representing pixels of a greyscale image; Figure 7 is a schematic block diagram of a digital camouflaging apparatus 15 according to the present invention; Figure 8 is a pseudocode of the digital camouflaging text encoder; Figure 9 is a pseudocode of the digital camouflaging text decoder; Figure 10 is a pseudocode of the spatial location program used in the encoding process shown in Figure 8; 20 Figure 11 illustrates an example of digital camouflaging of an alphanumeric text message in an image; and Figure 12 illustrates an entire English text paragraph which is camouflaged in the image of Figure 11. 25 Description of Preferred Embodiments In a preferred embodiment of the present invention, involving digital camouflaging, the invention exploits the data values present in various digital formats, such as for audio, image and video, for encoding and decoding of alphanumeric character strings. 30 Figure 1A illustrates a one dimensional data array containing integer values which may, for example, represent a digital audio signal (speech or music) WO 99/48006 PCT/SG98/00023 10 sampled under a conventional digital sound format. Figure 1B illustrates a two dimensional data array containing integer values which may, for example, represent a digital still image in GIF or JPEG format, or a video frame in MPEG format. 5 Figure 2 illustrates a typical digital sampled audio waveform (music) in WAV format, with amplitude (vertical axis) plotted against time (horizontal axis). Figure 3 illustrates a typical 8 bit greyscale image of 512 x 512 pixels. Digital camouflaging may be described as analogous to natural camouflaging, where the camouflaged object (for example a leaf, insect or reptile) 10 conceals itself completely into the surrounding environment. In the present invention, digital camouflaging conceals secondary data, such as an alphanumeric text string in integer form, in primary data, stoch as an image, by locating the spatial positions of integer values that match the text string values. Figure 4 illustrates a table of the ASCII character set. The digits at the left 15 of the table are the left digits of the decimal equivalent (0 - 127) of the character code, and the digits at the top of the table are the right digits of the character code. For example, the character code for "F" is 70, and the character code for "&" is 38. The commonly used alphanumeric codes for text strings are as follows: 48 to 57 represent the numeric values "0 - 9", 20 65 to 90 represent the upper case alphabet "A - Z", and 97 to 122 represent the lower case alphabet "a - z". The present invention may be applied to many different data formats for audio, image and video data. In the case of digital image and video camouflaging, a histogram distribution for most natural images, such as human 25 faces or landscapes, would span a wide range of greyscale levels. For example, an 8 - bit or 256 grey level image would contain image pixel values between 0 and 255. Figure 5 illustrates a typical image histogram. It can be seen that a wide spread of pixel values is available in the image for exact matching to alphanumeric character codes to be camouflaged. It can also be seen that a 30 large number of pixels at different spatial locations within the image have the same integer value. For example, the value 105 appears in approximately 2500 WO 99/48006 PCT/SG98/00023 11 pixels in the image. For each alphanumeric character of a text string to be encoded in an image, a search for an exact matching of this character to an image pixel is first performed. Once the pixel value is found, the address of the spatial location of 5 the pixel is stored in a key element. However, more than one pixel containing the same value as the alphanumeric character value is likely to arise. When this occurs, the encoding method will preferably select the first pixel that it finds. The address of this pixel will form an important part of key data that will later be used to decode the camouflaged alphanumeric character text string. Table 1 gives an 10 example of an alphanumeric text string "This is an Example" and the corresponding integer values. TABLE 1 Alphanumeric Character Integer Value T 84 h 104 1 105 s 115 space 32 1 105 s 115 space 32 a 97 n 110 space 32 E 69 x 120 a 97 m 109 P 112 I 108 e 101 15 Thus, "This is an Example" is equivalent to the data array [84 104 105 115 32 105 115 32 97 110 32 69 120 97 109 112 108 101]. Figures 6A and 6B show an example of a two-dimensional data array representing pixel values of a 4 x 8 pixel image. The terms (1, 1), (1, 2) ..... (4, 8) WO 99/48006 PCT/SG98/00023 12 are spatial location addresses of the pixels in the image. These addresses are stored as key elements when the alphanumeric character values are matched to pixel values during searching. For example, in the encoding process, the first alphanumeric character "T" with its integer value of 84 (shown in bold in Figure 5 6A) would match the values of pixels located at (3,1) and (2, 8). The address of the first location (3, 1) will be stored in a corresponding key element for later use in the complementary decoding process. There are some instances, however, where after searching through all the pixel values in an image, a particular alphanumeric character value does not 10 result in an exact match to its value. In these instances, the present invention will perform a new search for an adjusted value that is one pixel off either side of the character value. The address of the pixel containing the adjusted value is then stored as the key element. For example, if the first alphanumeric character of the text string is a "T", with an integer value of 84, when the search of the pixel values 15 does not locate a single value of 84, a new search would commence to locate values 83 or 85. These values correspond to alphanumeric characters of "S" and "U", respectively. In one embodiment, when the new search locates a pixel having a value of 83 or 85 it will first check whether that pixel has already been used. If so, and 20 providing that the same pixel values occur at other addresses, the encoding method will select another pixel having the value of 83 or 85. Once such a pixel has been found, the pixel value is overwritten with the character value of 84. If the adjusted value, being one pixel value off either side of the character value, cannot be found, then a second search commences to locate a pixel value 25 having a new adjusted value, of 2 pixel values off either side of the character value (ie. 82 and 86 in the present example). The matching and overwriting steps are then repeated as before. Referred to Figure 6B, and using the same example of alphanumeric text and image pixel values as in Figure 6A, the third and sixth alphanumeric 30 characters "i", which correspond to the value 105, will not find an exact match after searching the (4 x 8) image, as there is no such value existing. In this case, WO 99/48006 PCT/SG98/00023 13 the encoding method will search for a value that is an increment or decrement to that value. The method steps involved in "off-pixel" searching and overwriting are as follows: 5 1. Search for character value of 105 ("i") in image. 2. None found. 3. Increment or decrement value to 104 or 106. 4. None found for 106. 5. Two pixels found having value 104 (shown in bold in Figure 6B). 10 6. Check whether pixel has been used previously. 7. If yes, store second pixel spatial location address; overwrite pixel value in image with character value 105. 8. If no, store first pixel spatial location address; overwrite pixel value in image with character value 105. 15 Most natural images contain pixel values spanning over a wide range, thus allowing almost all alphanumeric character values to be exactly matched. After many tests with different natural images it has been found that the maximum number of off-pixel searches required is generally less than + 3 pixel values. Correlation analysis preformed on the same image before and after data 20 camouflaging using the present invention indicates that the images are often exactly identical. Even when there is no exact match of the pixel values to the alphanumeric character values, the minute change caused by overwriting just a few pixels in the image, for example from 104 to 105, has an insignificant degrading affect on image quality. Under these conditions, the correlation 25 coefficient between the two images still results, for all practical cases, in a value of substantially one. The complementary decoding process is relatively straight forward as it uses the unique key data generated from the spatial location addresses of pixels that match the alphanumeric characters composing the text string. The unique 30 key data provides the addresses of the spatial locations from which the camouflaged text string may be extracted from the image.

WO 99/48006 PCT/SG98/00023 14 The key data is unique to a specific text string and specific image. If the key data is applied to another image, the text string extracted will be just a random sequence of characters. For digital audio camouflaging, the dynamic range of values of audio 5 samples associated with music or speech may not be as wide as in the case for digital still/video images. The audio integer values may also be in a range that is somewhat different to the range of alphanumeric character values. A slightly different approach is therefore needed for camouflaging of alphanumeric text strings in digital audio data. 10 The present invention may accordingly include an adaptive statistical approach to first determine the dynamic range of the audio samples. The mean (ji) and the standard deviation (a) of the audio samples are determined and used to calculate a lower limit of the audio samples. For example, using the alphanumeric character code for "A", which corresponds to an integer value of 65, 15 as a reference, the alphanumeric character set may be adaptively shifted to coincide with the dynamic range of the audio samples. This adaptive statistical approach has proven very robust for embedding alphanumeric text strings into digital audio samples of music and speech. The method steps involved in the adaptive statistical range shifting process are preferably as follows: 20 Encoding 1. Store integer value of alphanumeric reference character, eg. 65 for "A". 2. Determine mean and standard deviation of audio samples. 3. Calculate lower limit of audio sample range, eg. by subtracting three times 25 the standard deviation from the mean (ie. t - 3a). 4. Obtain "data shifting offset" by subtracting reference character value from the lower limit of the audio sample range. 5. Add data shifting offset to each alphanumeric character. 6. Perform search, key element generation and overwriting steps as 30 described above in relation to image example. 7. Generate unique key data from key elements.

WO 99/48006 PCT/SG98/00023 15 Decoding 8. Extract audio sample values based on locations addressed in key elements of unique key data. 9. Reverse data shifting process, by subtracting data shifting offset from 5 values extracted, to obtain alphanumeric character values. Referring now to Figure 7, there is shown a schematic block diagram of a digital encoding/decoding apparatus according to an embodiment of the present invention. Data samples 10 are encoded with an alphanumeric character string 10 12 through a data encoding means 14. Encoded data samples 16 and unique key data 18, containing spatial location addresses, are obtained from the data encoding means 14. These two outputs 16, 18 are used in the reverse decoding means 20. The original alphanumeric character string 12 is extracted as 12A from the encoded data samples 16 using the location address contained in the 15 unique key data 18. The encoded data samples 16 remain unchanged or minimally changed as 16A by the decoding means 20. A pseudocode of the digital camouflaging text encoder is shown in Figure 8 and a pseudocode of the digital camouflaging text decoder is shown in Figure 9. A pseudocode of the spatial location program used in the encoding process is 20 shown in Figure 10. In these figures the term "unlabelled" is used to refer to the original data and the term "labelled" is used to refer to the encoded data. An example of the use of the present invention for camouflaging text labels in a digital image is illustrated in Figure 11. A text string containing 83 characters "This example shows an alphanumeric text message camouflaged into the image 25 of Lena." is camouflaged in the Lena image. The correlation coefficient equals exactly one, indicating that there is no difference between the labelled and unlabelled images. Moreover, an entire paragraph of English text containing 126 words and 784 characters, as shown in Figure 12, is camouflaged in the same Lena image and, once again, the correlation coefficient is equal to exactly one. 30 The present invention of camouflaging alphanumeric character text strings into digital data has many potential applications for resolving copyright protection WO 99/48006 PCT/SG98/00023 16 issues in the consumer sector or for secure transmission of messages in the commercial and service sectors. For example, data camouflaging may be incorporated into consumer electronic products, such as digital still/video cameras and, more recently, VCD and DVD players, to authenticate the true ownership of 5 intellectual property rights or product ownership. Another major consumer area for data camouflaging is that of preventing illegal copying and downloading of satellite images, music CD's, and tapes. In the commercial sector, copyright protection of multimedia data on the Internet needs also to be monitored closely, as there is a tremendous amount of 10 original data, in the form of music, image and video, illegally downloaded and redistributed without the consent of the true owners. The present invention can be used to address this problem, as well as providing for the secure transmission of messages in commercial and other operations. Similarly, in the service sector such as security and banking, secret text messages can be camouflaged in an 15 ordinary image or speech for secure transmission. Those skilled in the art would appreciate that various adaptations and modifications of the just described preferred embodiments may be configured without departing from the scope and spirit of the invention. Therefore, it is to be understood that, within the scope of the appended claims, the invention may be 20 practiced other than as specifically described herein.

Claims

1. A method of steganographic encoding including the steps of: (A) providing primary data containing a plurality of addressable first data 5 elements; (B) providing secondary data containing a plurality of second data elements; and (C) for each second data element: (Cl) searching for a first data element which matches content of 10 the second data element, and (C2) generating a key element including the address of the matching first data element when a match is found for the content of the second data element. 15

2. A method of steganographic encoding as defined in claim 1 wherein the first and second data elements are represented by integer values and wherein step (C) further includes: when a match is not found for the content of a second data element; (C3) producing an adjusted second data element by incrementing 20 or decrementing the integer value of the second data element, (C4) searching for a first data element which matches the integer value of the adjusted second data element, and (C5) generating a key element including the address of the matching first data element when a match is found for the adjusted second 25 data element, and replacing the content of the matching first data element with the integer value of the second data element prior to producing the adjusted second data element.

3. A method of steganographic encoding as defined in claim 2 wherein step 30 (C) further includes: when a match is not found for the content of the adjusted second WO 99/48006 PCT/SG98/00023 18 data element; (C6) producing a new adjusted second data element by incrementing or decrementing the adjusted second data element and repeating steps (C4) and (C5) for the new adjusted second data element. 5

4. A method of steganographic encoding as defined in claim 1 wherein, prior to step (C) the method includes: determining a range for the contents of the first data elements, determining a range for the contents of the second data elements, 10 comparing the range for the first data elements with the range for the second data elements, shifting the contents of the second data elements when the range for the second data elements falls outside of the range for the first data elements, such that the range for the second data elements falls substantially within the range for 15 the first data elements, and using the shifted second data elements as the second data elements in step (C).

5. A method of steganographic encoding as defined in claim 4 wherein the 20 step of determining a range for the contents of the first data elements includes: calculating a mean and standard deviation for the first data elements; and determining a lower limit for the first data elements based on the mean and standard deviation. 25

6. A method of steganographic encoding as defined in claim 4 wherein the step of determining a range for the contents of the second data elements includes: establishing as a reference a minimum value which can be attributed to the range of possible second data elements. 30

7. A method of steganographic encoding as defined in claim 6 wherein the WO 99/48006 PCT/SG98/00023 19 step of comparing includes calculating an offset value by subtracting the reference value from the lower limit.

8. A method of steganographic encoding as defined in claim 7 wherein the 5 step of shifting includes adding the offset value to the contents of each second data element.

9. A method of steganographic encoding as defined in claim 1, further including storing the key elements. 10

10. A method of steganographic encoding as defined in claim 7 further including storing the offset value with the key elements.

11. A method of steganographic encoding as defined in claim 1 wherein: 15 step (A) includes providing a digital representation of the content of each first data element, step (B) includes providing a digital representation of the content of each second data element, and step (Cl) includes comparing values of the digital representations of the 20 first and second data elements.

12. A method of steganographic encoding as defined in claim 1 wherein the secondary data includes a text message and each second data element includes an alphanumeric character. 25

13. A method of steganographic encoding as defined in claim 12 wherein the text message includes one or more of the following: a title, an artist, 30 a copyright holder, a body to which royalties should be paid, and WO 99/48006 PCT/SG98/00023 20 general terms for publisher distribution.

14. A method of steganographic encoding as defined in claim 2 further including storing the primary data or modified primary data produced by step (C5). 5

15. A method of steganographic encoding as defined in claim 1 wherein the primary data includes first data elements arranged in a two-dimensional array wherein each first data element defines a characteristic associated with a still image element. 10

16. A method of steganographic encoding as defined in claim 15 wherein the first data elements are obtained from a stream of data representing a digitised still image. 15

17. A method of steganographic encoding as defined in claim 16 wherein the digitised still image is obtained from a still digital camera.

18. A method of steganographic encoding as defined in claim 16 wherein the digitised still image is obtained from a computer game or other software. 20

19. A method of steganographic encoding as defined in claim 1 wherein the primary data includes first data elements arranged in a three-dimensional array wherein each first data element defines a characteristic associated with a motion video element. 25

20. A method of steganographic encoding as defined in claim 19 wherein the first data elements are obtained from a stream of data representing digitised motion video. 30

21. A method of steganographic encoding as defined in claim 20 wherein the digitised video is obtained from a Video Compact Disc player. WO 99/48006 PCT/SG98/00023 21

22. A method of steganographic encoding as defined in claim 20 wherein the digitised video is obtained from a Laser Disc player. 5

23. A method of steganographic encoding as defined in claim 20 wherein the digitised video is obtained from a digitised movie contained within a computer game or other software.

24. A method of steganographic encoding as defined in claim 20 wherein the 10 digitised video is obtained from a Digital Versatile Disc player.

25. A method of steganographic encoding as defined in claim 1 wherein the primary data includes first data elements arranged in a one-dimensional array wherein each first data element defines a characteristic associated with a digital 15 audio sample.

26. A method of steganographic encoding as defined in claim 25 wherein the digital audio samples are obtained from a stream of data representing digitised sound or music. 20

27. A method of steganographic encoding as defined in claim 25 wherein the digital audio samples are obtained from two streams of data representing two channels of digitised sound for stereo reproduction. 25

28. A method of steganographic encoding as defined in claim 25 wherein the digitised sound or music is obtained from a Compact Disc player.

29. A method of steganographic encoding as defined in claim 25 wherein the digitised sound or music is obtained from a Digital Audio Tape player. 30

30. A method of steganographic encoding as defined in claim 25 wherein the WO 99/48006 PCT/SG98/00023 22 digitised sound or music is obtained from a Laser Disc player.

31. A method of steganographic encoding as defined in claim 25 wherein the digitised sound or music is obtained from a Video Compact Disc player. 5

32. A method of steganographic encoding as defined in claim 1 wherein the number of first data elements is greater than the number of second data elements. 10

33. A method of steganographic encoding as defined in claim 1 wherein the address of each first data element is used only once for generating the key elements, whereby second data elements having equal contents are matched to first data elements having different addresses. 15

34. A method of steganographic decoding of secondary data including a plurality of second data elements, said secondary data being encoded in key elements in association with primary data, said method including the steps of: (A) providing said primary data containing a plurality of addressable first data elements; 20 (B) providing said key elements, each key element including an address of a first data element; and (C) for each key element, generating a said second data element by extracting the content of the addressed first data element. 25

35. A method of steganographic decoding as defined in claim 34 wherein the contents of the second data elements are shifted by an offset value further including: (D) providing said offset value, and (E) subtracting the offset value from the contents of the second data 30 elements. WO 99/48006 PCT/SG98/00023 23

36. A method of steganographic decoding as defined in claim 34 wherein each second data element includes an alphanumeric character.

37. An apparatus for steganographic encoding including: 5 (A) means for providing primary data containing a plurality of addressable first data elements; (B) means for providing secondary data containing a plurality of second data elements; (C) means for searching, for each second data element, a first data 10 element which matches content of the second data element, and (D) means for generating a key element including the address of the matching first data element when a match is found for the content of the second data element. 15

38. An apparatus for steganographic encoding as defined in claim 37 wherein the first and second data elements are represented by integer values, said apparatus further including: (E) means for producing an adjusted second data element by incrementing or decrementing the integer value of the second data element when 20 a match is not found for the content of a secondary data element, (F) means for searching for a first data element which matches the integer value of the adjusted second data element, (G) means for generating a key element including the address of the matching first data element when a match is found for the adjusted second data 25 element, and (H) means for replacing the content of the matching first data element with the integer value of the second data element prior to producing the adjusted second data element. 30

39. An apparatus for steganographic encoding as defined in claim 37 further including, means for producing a new adjusted data element by incrementing or WO 99/48006 PCT/SG98/00023 24 decrementing the adjusted second data element when a match is not found for the adjusted second data element.

40. An apparatus for steganographic encoding as defined in claim 37 further 5 including: means for determining a range for the contents of the first data elements, means for determining a range for the contents of the second data elements, means for comparing the range for the first data elements with the range 10 for the second data elements, means for shifting the contents of the second data elements when the range for the second data elements falls outside of the range for the first data elements, such that the range for the second data elements falls substantially within the range for the first data elements, and 15 means for using the shifted second data elements as the second data elements.

41. An apparatus for steganographic decoding of secondary data including a plurality of second data elements, said secondary data being encoded in key 20 elements in association with primary data, said apparatus including: (A) means for providing said primary data containing a plurality of addressable first data elements; (B) means for providing said key elements, each key element including an address of a first data element; and 25 (C) means for generating a second data element for each key element by extracting the content of the addressed first data element.

42. A method of cryptographic encoding including the steps of: (A) providing primary data containing a plurality of addressable first data 30 elements; (B) providing secondary data containing a plurality of second data WO 99/48006 PCT/SG98/00023 25 elements; and (C) for each second data element: (Cl) searching for a first data element which matches content of the second data element, and 5 (C2) generating a key element including the address of the matching first data element when a match is found for the content of the second data element.

43. A method of cryptographic decoding of secondary data including a plurality 10 of second data elements, said secondary data being encoded in key elements in association with primary data, said method including the steps of: (A) providing said primary data containing a plurality of addressable first data elements; (B) providing said key elements, each key element including an address 15 of a first data element; and (C) for each key element, generating a second data element by extracting the content of the addressed first data element.

44. An apparatus for cryptographic encoding including: 20 (A) means for providing primary data containing a plurality of addressable first data elements; (B) means for providing secondary data containing a plurality of second data elements; (C) means for searching for each second data element a first data 25 element which matches content of the second data element, and (D) means for generating a key element including the address of the matching first data element when a match is found for the content of the second data element. 30

45. An apparatus for cryptographic decoding of secondary data including a plurality of second data elements, said secondary data being encoded in key WO 99/48006 PCT/SG98/00023 26 elements in association with primary data, said apparatus including: (A) means for providing said primary data containing a plurality of addressable first data elements; (B) means for providing said key elements, each key element including 5 an address of a first data element; and (C) means for generating a second data element for each key element by extracting the content of the addressed first data element.