Patent Application
20060020806
1. Field of the Invention
This invention relates to a method for imprinting identification information (ID) into a digital content and for reading that information.
2. Background of the Related Art
The information superhighway was advocated in the United States in 1991, and since then distribution of information over networks as represented by the Internet has been forming a new society base. In this new network society, secure encryption and authentication are desired in such fields as electronic commerce because such fields are concerned with safety.
On the other hand, one of the principles of the Internet is the free distribution of digital contents such as pictures, animation and music (hereinafter collectively referred to as contents). Presently, even for valuable content, such as cultural works, illegal copies can be easily made and distributed. Collecting fees for using contents on the Internet, preventing illegal reproduction or modification, and protecting copyrights are serious problems that need to be addressed and solved. These issues are extremely important for the mutual development of a network society and culture.
It is therefore desired to design a general approach to trace illegal copies of digital contents.
The above references are incorporated by reference herein where appropriate for appropriate teachings of additional or alternative details, features and/or technical background.
An object of the invention is to solve at least the above problems and/or disadvantages and to provide at least the advantages described hereinafter.
Another object of this invention to provide an identification (ID) imprinting method applicable to existing contents.
It is a further object of this invention to provide an ID imprinting method applicable to a content having no reserved areas or areas for remarks that do not play any role in the content.
It is still another object of this invention to provide an ID imprinting method which does not introduce substantial degradation of the content quality when an ID information is imprinted.
It is yet another object of this invention to provide an ID imprinting method for embedding an ID information that can be easily detected.
It is yet another object of this invention to provide an ID readout method to easily detect and interpret the ID information imprinted in the content.
A method according to an embodiment of the present invention comprises loading a content into an information terminal where the content is used and imprinting an ID information associated to the information terminal or its user into a predetermined location in a perceivable portion of the loaded content. (A content may be any collection of digital data, and may be in the form of a sequence of data values. A perceivable portion contains data that play a role in the content, rather than reserved areas or areas for remarks that do not play any role in the content.)
The content is first loaded into an information terminal. Subsequently, an ID information is imprinted into a predetermined location of the content. A user who reproduces illegal copies of the content is identified with the ID information imprinted therein. Since the ID information is imprinted in a predetermined location, no string search is necessary. This method is applicable to existing contents, since it requires no special data blocks beforehand.
In another aspect of the invention, an ID information is imprinted in the form of spatial frequency information into the entire content loaded into an information terminal. “Spatial frequency information” is information relative to a spatial frequency in any sense. In this aspect, the ID information is converted into spatial frequency information via, for instance, an inverse orthogonal transformation so as to be reflected in the content data. The inverse orthogonal transformation may be an inverse fast Fourier transform (IFFT) or an inverse discrete cosine transform (IDCT). This method is also applicable to existing contents.
According to the ID reading process of this invention, a content is first obtained for instance via a network, and an ID information is read from a predetermined location thereof. The ID information is uniquely associated with an information terminal or its user. In another aspect, spatial frequency information is extracted from the obtained content, and then supplied for an orthogonal transformation. Through the transformation, the ID information imprinted in the content is restored. An orthogonal transformation may be a fast Fourier transform (FFT), a discrete cosine transform (DCT), and so forth.
To achieve at least the above objects in a whole or in part, there is provided a method for distributing a digital content on a computer network that includes receiving via the computer network a request for a digital content from an information terminal, the request identifying the information terminal or a user thereof, obtaining an encoded version of the digital content, wherein the encoded version of the digital content is generated by performing encryption and compression on the original digital content, generating a decoding program such that ID information representing the information terminal or user is embedded therein, and wherein the decoding program is programmed to perform the function of decrypting and decompressing the encoded digital content at the information terminal to produce a decoded digital content having the ID information imprinted therein, transmitting via the computer network the decoding program to the information terminal, and transmitting via the computer network the encoded digital content to the information terminal in response to the request.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and advantages of the invention may be realized and attained as particularly pointed out in the appended claims.
The invention will be described in detail with reference to the following drawings in which like reference numerals refer to like elements wherein:
1 is added to the luminance of respective pixels to change the state shown in
The present invention may be applied to a network system comprising a server 2 and client devices connected on a network 9, as shown in
The server 2 supplies a content to the client devices so that ID imprinting is carried out on the client side. Here, as an example, the PC 4 is provided with an imprinting function.
The viewer 12 received in the PC 4 may now be used to decrypt and display the received content. As shown in
As a measure for preventing use of content before ID imprinting, for instance, a memory area in the PC 4 that stores a content without an ID imprinted therein is protected by the ID imprinter 18 so that reading of such a content is prevented. Specifically, the system is designed so as to be interrupted or reset if a read access is made to the memory area containing a content without an ID. Once an ID is imprinted, this protection is removed, enabling the image to be used as desired.
As shown in
In operation, the user of the PC 4 requests the server 2 run by the content manager to transmit a content. The content is encrypted on the server 2 and then sent with the viewer 12 to the PC 4 via a network. The communication section 10 of the PC 4 receives the transmitted content, and forwards it to the viewer 12, which has been received from the network. Within the viewer 12, the image decoder 14 decodes the content, and forwards it to the ID imprinter 18. The ID reader 30 in the ID imprinter 18 reads the ID from the ID holder 16 and supplies it to the combiner 34. The decoded image reader 32 reads the decoded image, and forwards it to the combiner 34. Having received the ID and the decoded image, the combiner 34 replaces the LSBs of the luminance in the aforementioned manner to thereby imprint the ID onto the image. The image having the ID is displayed on the display 24. The ID imprinted image may also be supplied to the memory 26. If a subsequent unauthorized attempt is made to modify or reproduce the ID-imprinted image stored in the memory 26, copies of such modified or reproduced image will carry the ID information imprinted in the image stored in the memory 26. It is therefore possible to identify the party making the unauthorized copies.
In this embodiment, the ID information is represented as a signal in the frequency domain. When imprinting such an ID, an inverse orthogonal transform is applied to the frequency signal representing the ID information to generate a bit pattern in the content domain, which is then imprinted in the digital content. In this specification, the term “content domain” is used to denote the domain representing the data in the digital content, while the term “frequency domain” is used to denote a mapping of the content domain through an orthogonal transform. When the content is a two-dimensional image (an example used in the illustration below), the content domain is a two-dimensional space domain, and the corresponding frequency domain is a two-dimensional spatial frequency domain. When the content is audio, the content domain may be a time domain and the frequency domain may be a one-dimensional frequency domain.
Specific embodiments of the present invention for imprinting ID information have been described. Many variations of the embodiments are possible, some of which are described below.
First, although a content is distributed via a network in the above-described embodiments, the content may also be distributed by storing it in a medium such as a CD-ROM or a floppy disc and loading it onto a PC. The embodiments described above are applicable to such other methods of content distribution.
Second, although a still image is used in the above-described embodiments as an example of a digital content, the methods may be applied to other types of digital content, such as motion images (e.g. video) or audio content. For audio content, the image decoder 14, the display controller 20, and the display 24 in
Third, an ID is not necessarily stored in the LSBs of a content. Any bits of quantified data may store the ID as long as the effects on the perceived quality of the content are insignificant. It should be noted that even perceptible imprinting may be employed as a visual watermark.
Fourth, although an ID is imprinted into a lower bit irrespective of upper bits in the aforementioned embodiment, an offset may be given to a lower bit such that the whole data including upper bits contains the ID.
Since this method can prevent direct exposure of an ID unlike imprinting it in the lower bits, security is increased. Another advantage is that data other than “0” and “1”, such as “2”, is also imprintable. Although modulo 3 arithmetic is mentioned here, modulo arithmetic based on other numbers may be used. Any other mathematical, boolean algebraic or cryptographic approach may be employed.
Fifth, in the aforementioned embodiments, the combiner 34 (
Sixth, in the embodiment shown in
Seventh, in the embodiment shown in
Eighth, although the ID information for one user is represented by one point in the frequency domain (
Ninth, in a frequency domain representation such as that shown in
Methods for imprinting ID information have been described. Methods for reading imprinted ID information will be described next.
If a content is illegally reproduced or modified (hereinafter referred to as an illegal action), it is desired that the unauthorized offender be identified. This can be achieved by reading the ID information imprinted into the content. A device for reading imprinted ID information (hereinafter referred to as a detector) may be provided anywhere in a network. A proxy server, for instance, may be equipped with such a detector.
In this embodiment, the ID reader 64 extracts information from a predetermined location, for instance, the LSBs at a leading part of a data sequence of the obtained content, and reconstructs the ID based on the extracted data. If this process does not result in any ID information that meaningfully identifies a user, then the content is judged to be original, i.e., having no user ID information imprinted. On the other hand, if a content with a user's ID imprinted therein is found on a network, the user identified by the imprinted ID may have illegally distributed the content. Based on the ID, the possible illegal action is traced.
The operation here is a reverse operation of that shown in FIGS. 6 to 8. The LSBs, which represent the imprinted ID information, are first extracted (
This method is advantageous in that it does not require comparing the suspect content and the original content in order to detect the ID.
The above-described methods for reading imprinted ID information may have many variations. Each variation of the ID imprinting method described earlier in this specification may have a corresponding variation of the ID reading method. For example, the ID can be read in cases where an offset has been added to a lower bit using a method such as the one described earlier with reference to
In addition, although the detector is connected to a network in the above embodiments, it may be an off-line, stand alone type if it checks only contents stored in storage media.
Moreover, in the described embodiments, the ID imprinting is carried out at the information terminals where the content is used, i.e. at the user end. It will be apparent to a skilled artisan, however, that the various methods described herein for imprinting ID information in a content are equally applicable to a content distribution scheme in which ID imprinting is carried out at the content provider end.
The embodiments above have generally been described without regard to types and formats of contents. Also, as a processing to be performed by the server 2 before sending the content to the PC 4, there has generally been no explicit distinction made between different forms of encoding, such as encryption and compression. For example, how the encoding and decoding of content to be done by the viewer 12 in the PC4 corresponds to the decryption or decompression is not specified and constraints have not been described. In what is to follow, more specific processing examples are described in which the ID information is imprinted onto a content in accordance with the type of the content and a processing prior to the distribution of the content.
For the sake of convenience, some terms will be defined as follows. Hereinbelow, “encryption” refers generally to a process by which content is encoded to prevent the use of the content by third parties. As one non-limiting example, scrambling or the like that utilizes XOR operation can be used as an encryption proceess. In addition, generally any known or subsequently developed method suitable for the encryption of a particular form of content may be used. Similarily, a process for restoring the encrypted content to a utilizable state is referred to as “decryption”.
In contrast to encryption, a process for reducing the amount of data to suit, for example, the distribution of content via a network is referred to as “compression” in what follows. For compression techniques, generally any known or subsequently developed techniques may be used. Non-limiting examples of known techniques include the JPEG scheme used for still images and the MPEG scheme used for motion images. Another known compression technique is the pixel-based image matching technique (hereinafter referred to as “CPF matching”) using a Critical Point Filter as disclosed in U.S. Pat. No. 6,018,592, owned by the same assignee as this present patent application. CPF matching is particularly suitable for use in the compression of motion images or the like. The process by which the compressed contents are restored to the original ones so as to utilize them is called “decompression”.
The general terms above have been defined for convenience. However, it is to be noted that encryption and compression as well as decryption and decompression are mutually overlapping concepts in the area of encoding and therefore the terms may be sometimes used interchangeably in the present patent application as the case may be. In addition, there are cases hereinbelow where some technical terms in the field of encoding may be used as broader concepts that include encryption and compression and also some technical terms in the field of decoding are used as broader concepts that include decryption and decompression. In the more specific examples described below, it is generally assumed that the contents to be distributed from the server 2 to the PC 4 are first compressed and are then encrypted.
Here, in the viewer 12, the image decoder 14 both decrypts and decompresses the compressed and encrypted content C″ so as to generate the decoded content C and, in so doing, the ID imprinter 18 may embed the ID information ((A) in
In the former case, the fully decoded content alone is subject to ID information embedding, and, as such, the processing for embedding the ID information is carried out with relative ease. This is because there will be no concern about ensuring that the embedded ID information is inherited without fail from the compressed content C′ to the decoded content as is needed in the latter case. On the other hand, in the latter case, the illegitimate distribution of the compressed content C′ can be deterred and/or tracked. Furthermore, in the latter case, the ID information embedded in the compressed content C′ is preferably also inherited to the decoded content C, thus achieving advantageous effects, without separately embedding the ID information into the decoded content C, such as preventing the illegal circulation of both the compressed content C′ and the decoded content C.
It will be understood that, in the latter example, the ID imprinter 18 may directly embed the ID information respectively into both the compressed content C′ and the decoded content C. In such a case, the ID information is reliably embedded into both the contents, thus assuring a higher level of security.
Hereinbelow, the types of contents will be described in more specific form and specific examples in which the ID information is embedded into the compressed content C′ and/or the decoded content C will be described.
(1) Still Images as an Example
In this example, the compressed and encrypted content C″ is a content produced after a still image (hereinafter referred to as original still image) is subjected to the encryption after it has already been compressed in the JPEG format in the server 2.
(i) Embedding into Decoded Content C
A decoded content C obtained after the compressed and encrypted content C″ has been subjected to a decryption processing and a decompression processing is a still image which is identical or approximate to the original still image. In order to imprint or embed the ID information into this still image, the aforementioned processing techniques, for example, where the LSB of a luminance value is replaced by a bit pattern containing the ID information for a predetermined number of pixels, can be utilized. Further, the ID information can be embedded, in the form of spatial frequency information, into the entire content. The ID information may also be detected using the aforementioned methods.
(ii) Embedding into Compressed Content C′
The compressed content C′ is JPEG data obtained by performing JPEG compression on an original still image. In this case, if the ID is of n bits, it is preferred that the ID information be embedded in a manner such that, for data corresponding to a predetermined block, for example, a leading block, among the compressed data, the LSB of each of n DCT coefficients of a luminance value among all DCT coefficients thereof is replaced by each of the respective bits of the ID. An object to which the ID information is to be embedded may be DCT coefficients for the color components. In order to detect the ID, the LSB of such DCT coefficients is referred to.
In this example, it is desirable that the ID information be embedded in such a position, in the data of compressed content C′, as to both maintain the image quality and disable the removal of the ID information by third parties. That is, when the ID information is solely embedded in the DCT coefficients of high-frequency component, a third party might deliberately perform a decompression processing in which the high-frequency components of the compressed content C′ are discarded to acquire images having no ID information and without much degradation in the quality of image in the decoded content C. On the other hand, if the ID information is embedded in the DCT coefficients corresponding to low-frequency components of the compressed content C′, it is highly probable that a larger difference from the original image results. Thus, it is desirable that the location at which the ID information is to be embedded is determined by taking these facts into account. For example, a location or method at which the embedding in low-frequency component can be achieved as much as possible to the extent that the deviation from the original image does not affect human perception is preferably determined through experiments or the like.
It will be understood that the ID information may be embedded directly into both the compressed content C′, namely, JPEG-compressed data of an original still image and the decoded data C, namely, a decoded still image content by combining the above two methods.
(2) Motion Images as Another Example
In this example, the compressed and encrypted content C″ is the result after a motion image content C, composed of a plurality of image frames, (hereinafter referred to as original motion images) are compressed using the CPF matching processing and then subjected to the encryption processing in the server 2. The compression using the CPF matching processing is such that the CPF matching is carried out between or among key frames extracted at intervals of, for example, 0.5 seconds or the like from motion images of 30 frames per second so as to generate a corresponding-point information file, and then the original motion image data, composed of a plurality of image frames, are replaced by key frame data and the corresponding point information so as to reduce the data amount. The compressed data are uncompressed in a manner such that the key frames are interpolated using the corresponding-point information file and then frames corresponding to original motion images, other than the key frames, are generated as intermediate images. Thereby, the original motion images or those approximate thereto can be produced. It is to be noted that in order to further increase the data compression rate of the entire content, the key frames themselves are generally subjected to intra-frame compression, such as JPEG. Here, as a unit of image, “frame” includes a concept such as “field” which may be treated similarly to the frame, and no distinction is made between the two in this specification.
(i) Embedding into Decoded Content C
A decoded content C obtained after a compressed and encrypted content C″ has been subjected to a decryption and a decompression processing is made up of motion images which are identical or approximate to the original motion images. These motion images are a set of still images. Thus, the ID information can be embedded in one or more predetermined image frames in the decoded content C by using the aforementioned method for embedding the ID information into the still image. This embedding can occur either in frames that were key frames during compression or in frames that have been generated during decompression (intermediate frames).
(ii) Embedding into Compressed Content C′
The compressed content C′ is a content for which the original motion image data are compressed using the CPF matching technique. The contents thereof include data on key frames compressed in JPEG format (intra-frame compression) and a file that contains information on corresponding points among the key frames (interframe compression).
The ID information can be embedded in the JPEG-compressed data of the key frames by using the aforementioned method similar to that for embedding the ID information in the compressed content C′ where the contents are still images, for example.
The ID information may also be embedded in a corresponding-point information file. Here, if in a key frame sequence the leading frame is denoted by a first key frame and a frame subsequent to the first key frame is denoted by a second key frame, then an example is considered herein where the ID information is embedded in the information file of corresponding points between the first key frame and the second key frame. The corresponding-point information file into which the ID information is embedded is not limited thereto, and the ID information may be embedded in an information file of corresponding points of a single key frame or among a plurality of key frames.
How each pixel in the first key frame corresponds to which pixel of the second key frame is described in the corresponding-point information file. The ID information can be embedded, for example, in a manner such that the destination pixel in the second key frame to which the leading pixel of the first key frame corresponds is replaced with a pixel specified by the ID information. That is, when the ID of user A is expressed in an n-bit binary number k, the destination to which the leading pixel of the first key frame corresponds to is replaced by the k-th pixel from the leading pixel of the second key frame. The ID information can be detected by checking which pixel of the second key frame is corresponded to the leading pixel of the first key frame.
The present invention has been described by way of specific examples where the digital content is either a still image or motion images. However, the type of content is not limited thereto, and the present invention may be applied to, for example, audio content and various other forms of digital content as well.
The foregoing embodiments and advantages are merely exemplary and are not to be construed as limiting the present invention. The present teaching can be readily applied to other types of apparatuses. The description of the present invention is intended to be illustrative, and not to limit the scope of the claims. Many alternatives, modifications, and variations will be apparent to those skilled in the art. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents but also equivalent structures.
| Number | Date | Country | Kind |
|---|---|---|---|
| 8-296830 | Nov 1996 | JP | national |
| 9-282468 | Sep 1997 | JP | national |
This application is a continuation-in-part of U.S. patent application Ser. No. 09/304,644, filed May 4, 1999, which is a continuation of PCT Application No. PCT/US97/20309, filed Nov. 6, 1997, which claims priority from Japanese Patent Applications 8-296830, filed Nov. 8, 1996 and 9-282468, filed Sep. 9, 1997. Incorporation by reference of the entire disclosure of the prior applications are considered as being part of the disclosure of the accompanying application and is hereby incorporated by reference therein.
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/US97/20309 | Nov 1997 | US |
| Child | 09304644 | May 1999 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 09304644 | May 1999 | US |
| Child | 11154985 | Jun 2005 | US |
