Electronic watermark embedding apparatus and method and electronic watermark examining apparatus and method

1. An electronic watermark embedding apparatus comprising:

input image storage means for storing a pixel matrix G with n rows and n columns of an input image;

eigenvalue-and-vector calculating means for acquiring eigenvalues s 1 to s n (s 1 >s 2 > . . . >s n ) and eigenvectors p 1 to p n of the pixel matrix G read from the input image storage means;

eigenvalue correcting means for selecting at least one pair of eigenvalues (s i ,s j ) having different values from the acquired eigenvalues s 1 to s n and correcting the acquired eigenvalues s 1 to s n such that the values of the selected pair can be equal;

means for creating a Jordan-normal-form matrix S w having the corrected eigenvalues as diagonal elements and 1 as one of the nondiagonal elements corresponding to the selected pair;

means for outputting the acquired eigenvectors p 1 to p n as electronic watermark key information;

means for performing the inverse transform of diagonalization computing on the created matrix S w by using the acquired eigenvectors p 1 to p n to create a pixel matrix G w with n rows and n columns thereby and outputting it as an electronic-watermark-embedded image of the input image.

2. The electronic watermark embedding apparatus according to claim 1 , wherein the eigenvalue correcting means replaces one value of the selected pair by the other value to correct the acquired eigenvalues s 1 to s n such that the selected pair can have equal values.

3. The electronic watermark embedding apparatus according to claim 1 , further comprising:

original image storage means for storing an original image which is to be the target for embedding an electronic watermark;

means for cutting out a partial image of n×n pixels from the original image and storing a pixel matrix of the partial image in the input image storage means as the pixel matrix G with n rows and n columns of the input image; and

means for replacing the partial image in the original image by the electronic-watermark-embedded image of the input image to create an electronic-watermark-embedded image of the original image and outputting it.

4. The electronic watermark embedding apparatus according to claim 1 , wherein:

if the pixel matrix G is an irregular matrix of a rank r (r<n), the eigenvalue-and-vector calculating means determines one or more positive values, each of which is lower than the eigenvalue s r , as the eigenvalues s (r+1) to s n of the pixel matrix G and determines one or more fundamental vectors as the eigenvector of the determined eigenvalues s (r+1) to s n .

5. An electronic watermark examining apparatus comprising:

examination target image storage mean for storing a pixel matrix G k with n rows and n columns of an examination target image;

key information storage means for storing eigenvectors p 1 to p n of a pixel matrix G with n rows and n columns of an input image as electronic watermark key information;

means for performing diagonalization computing on the pixel matrix G k , which is read from the examination target image storage means, by using the eigenvectors p 1 to p n to create a matrix S k with n rows and n columns; and

determining means for determining whether the created matrix S k corresponds to a Jordan-normal-form matrix or not, and, if so, determines the examination target image is an electronic-watermark-embedded image of the input image and outputting the determination result.

6. The electronic watermark examining apparatus according to claim 5 , wherein:

the key information storage means further stores eigenvalues s 1 to s n (s 1 >s 2 > . . . >s n ) of the pixel matrix G with n rows and n columns of the input image as electronic watermark key information; and

the determining means determines whether the created matrix S k corresponds to a Jordan-normal-form matrix S w with n rows and n columns having m (1≦m<n) eigenvalues being one of the eigenvalues s 1 to s n and the remaining (n−m) eigenvalues being a multiple root with one of the m eigenvalues or not, and, if so, determines that the examination target image is an electronic-watermark-embedded image of the input image.

7. The electronic watermark examining apparatus according to claim 5 , wherein:

the key information storage means further stores a Jordan-normal-form matrix S w as electronic watermark key information;

the determining means determines whether the created matrix S k corresponds to the Jordan-normal-form matrix S w or not, and, if so, determines that the examination target image is an electronic-watermark-embedded image of the input image; and

the Jordan normal form S w is a Jordan-normal-form matrix acquired by selecting at least one pair of eigenvalues having different values from eigenvalues s 1 to s n (s 1 >s 2 > . . . >s n ) of the pixel matrix G, correcting the acquired eigenvalues s 1 to s n such that the selected pair can have equal values, setting the corrected eigenvalues as diagonal elements, and setting 1 as one of the nondiagonal elements corresponding to the selected pair.

8. An electronic watermark embedding method comprising:

the step of reading a pixel matrix G with n rows and n columns of an input image from input image storage means storing the pixel matrix G;

an eigenvalue and vector calculating step of acquiring the eigenvalues s 1 to s n (s 1 >s 2 > . . . >s n ) and eigenvectors p 1 to p n of the read pixel matrix G;

an eigenvalue correcting step of selecting at least one pair of eigenvalues (s i ,s j ) having different values from the acquired eigenvalues s 1 to s n and correcting the acquired eigenvalues s 1 to s n such that the selected pair can have equal values;

the step of creating a Jordan-normal-form matrix S w having the corrected eigenvalues as diagonal elements and having 1 as one of the nondiagonal elements corresponding to the selected pair;

the step of outputting the acquired eigenvectors p 1 to p n as electronic watermark key information; and

the step of performing the inverse transform of diagonalization computing on the created matrix S w by using the acquired eigenvectors p 1 to p n to create a pixel matrix G w with n rows and n columns thereby and outputting it as an electronic-watermark-embedded image of the input image.

9. A program, stored on a non-transitory computer readable medium, for causing a computer to execute the electronic watermark embedding method according to claim 8 .

10. The electronic watermark examining method comprising the steps of:

reading a pixel matrix G k with n rows and n columns of an examination target image from examination target image storage means storing the pixel matrix G k ;

reading eigenvectors p 1 to p n of the pixel matrix G with n rows and n columns of the input image as electronic watermark key information from key information storage means storing eigenvectors p 1 to p n ;

performing diagonalization computing on the read pixel matrix G k by using the read eigenvectors p 1 to p n to create a matrix S k with n rows and n columns;

determining whether the created matrix S k corresponds to a Jordan-normal-form matrix or not; and

if the created matrix S k corresponds to a Jordan-normal-form matrix, determining that the examination target image is an electronic-watermark-embedded image of the input image and outputting the determination result.

11. A program, stored on a non-transitory computer readable medium, for causing a computer to execute the electronic watermark examining method according to claim 10 .