Electrical computers and digital processing systems: support – Multiple computer communication using cryptography – Particular communication authentication technique
Reexamination Certificate
1997-12-24
2001-06-19
Swann, Tod R. (Department: 2767)
Electrical computers and digital processing systems: support
Multiple computer communication using cryptography
Particular communication authentication technique
C382S232000, C380S054000
Reexamination Certificate
active
06249870
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a data hiding method and a data extraction method to embed owner or copyright information (message information) into media information such as digital images, digital videos, or digital audio in an unperceptive manner. Particularly, it relates to a data hiding method to embed media information in an unperceptive manner while controlling the embedding operation for the information with a statistical examination. It also relates to a data extraction method to determine with the statistical examination whether or not the media information is embedded, and to properly retrieve the embedded information based on the determination.
BACKGROUND ART
The following technique has been known as a data hiding method with a statistical approach. First, two pixel point arrays, hereinafter called {an} and {bn}, are selected from the image data. Each respective pixel point array is composed of a number “n” of pixel points. Then, an embedding operation is performed by adding a fixed value c to the value v(an) of n pixels in point array {an}, while subtracting the fixed value c from the value v (bn) of “n” pixels in the other point array {bn}. Thus at each pixel “n”, the value of the new array vO(an) and vO(bn) are given by
v
0
(an)=v (an)+c
v
0
(bn)=v (bn)−c [Equation 1]
Whether or not embedding is conducted in image data is determined by calculating the average difference between the values of n pixels in both point arrays, and based on its result, as represented in the following formula.
1
N
∑
n
=
1
N
&AutoLeftMatch;
(
v
′
(
a
n
)
)
-
v
′
(
b
n
)
)
[
Equation
2
]
That is, when an average is calculated for the differences between values of pixels when the number “n” is high enough such that a statistical property appears, it is expected that the average converges to zero if no addition (embedding) is conducted. On the other hand, if the addition (embedding) is conducted, it is expected that the average becomes a fixed value of 2
c.
Accordingly, whether or not the embedding operation is performed is determined on the basis of a fact whether the average is closer to 0 or 2
c,
within a set threshold value.
The information to be embedded is one bit, and the positions of the two selected pixel point arrays are confidential. These are known only to the person who performs the embedding (addition) operation. Since the average 2
c
cannot be retrieved by a person other than the person who performs the embedding operation unless the positions of pixel point arrays can be identified, the fact that this value can be extracted proves that the person is the owner of the data.
However, a problem with this conventional technique that uses the statistical approach, lies in that the identification of who is the owner of data can be easily made unclear by conducting a new embedding operation. That is, even if a third party does not manipulate the pixel value in the pixel point array known only to the owner, the average 2
c
can be newly produced in two other suitable pixel point arrays by selecting such pixel point arrays, and conducting the embedding operation on their pixel values. Therefore, when the third party who performed such operation falsely claims that he/she himself/herself is the owner based on the average that is calculated from the point array identified by him or her, the conventional technique cannot prove who is the genuine owner.
Another problem lies in that sufficient consideration is not given to the embedding operation and deterioration of image quality caused by it. Even if the same value c is added or subtracted, significant difference may occur in deterioration of visual image quality because of certain characteristics of the image quality. Therefore, it is preferable to adaptively change the amount of embedding, (c), depending on the characteristics of image quality. That is, it is preferable to adaptively select the fixed value c and the number of pixels N in the pixel point arrays.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a data hiding method which enables it to be used to properly prove who is the genuine owner.
Another object of the present invention is to inhibit deterioration of image quality by adaptively determining the value of c to be added (embedded) to be dependent upon the characteristic values (such as pixel values in the medium information) on which the embedding processing is conducted.
A feature of the present invention is to use a key for selecting a pixel point array. This prevents false authentication of the owner. Another feature is to adaptively determine the size of selected pixel point array (N), and the magnitude of statistic amount (c) to be manipulated for the characteristic values. In retrieving the data, a calculation is performed to determine whether or not information is embedded in a region. This is used to determine whether or not the arithmetic operation is conducted on the characteristic values in that region. This also enables one to embed multi-bit information (author, distribution ID, and the like) in an image in an unperceptive manner, and to retrieve the embedded information from the image for claiming the ownership of the image or for identifying the distribution channel.
More specifically, an aspect of the present invention is a data hiding method for embedding one or more messages into data comprising the steps of: obtaining a secondary key by inputting a key having a predetermined value to a specified function; determining hiding positions in which the message will be embedded according to the obtained secondary key, and selecting one of many hiding functions according to the secondary key so that the selected hiding function is used for each of the determined hiding positions; and embedding the message into each of the determined hiding positions according to the selected hiding function corresponding to the hiding position.
Another aspect of the present invention is a data extraction method for extracting one or more messages from data in which each message is embedded in the image data. The method comprising the steps of: obtaining a secondary key by inputting a key having a predetermined value to a specified function; determining hiding positions in which the message is embedded according to the obtained secondary key, and selecting one of many hiding functions according to the secondary key so that the selected hiding function is used for each of the determined hiding positions; inputting information in the determined position to a detection basic function which is identified according to the determined hiding function; inputting an output of the detection basic function to a detection function; and determining whether the message is embedded according to an output of the detection function.
REFERENCES:
patent: 5613004 (1997-03-01), Cooperman et al.
patent: 5636292 (1997-06-01), Rhoads
patent: 5689587 (1997-11-01), Bender et al.
patent: 5721788 (1998-02-01), Powell et al.
patent: 5822436 (1998-10-01), Rhoads
patent: 5825892 (1998-10-01), Braudaway et al.
patent: 5875249 (1999-02-01), Mintzer et al.
patent: 5889868 (1999-03-01), Moskowitz et al.
patent: 5905800 (1999-05-01), Moskowitz et al.
patent: 5907618 (1999-05-01), Gennaro et al.
patent: 5907619 (1999-05-01), Davis
patent: 851392 A2 (1998-07-01), None
patent: 851391 A2 (1998-07-01), None
patent: 852441 A1 (1998-07-01), None
“Embedding Parametric Digital Signatures in Images”, by Adrian G. Bors, et al., pp. 1701-1704.
“Embedding Robust Labels into Images for Copyright Protection”, by Jian Zhao, et al., pp. 243-251.
Kobayashi Seiji
Koide Akio
Morimoto Norishige
Shimizu Shuichi
Hertzberg Louis P.
International Business Machines - Corporation
Kabakoff Steve
Scully Scott Murphy & Presser
Swann Tod R.
LandOfFree
Data hiding method and data extraction method using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data hiding method and data extraction method using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data hiding method and data extraction method using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2480979