Information embedding method, apparatus and recording medium

Details Information embedding method, apparatus and recording medium Information embedding method, apparatus and recording medium

1999-07-28

2003-07-08

Patel, Jayanti K. (Department: 2625)

Image analysis

Color image processing

Pattern recognition or classification using color

C382S164000, C382S232000, C382S234000, C380S201000, C380S210000, C713S176000, C713S179000

Reexamination Certificate

active

06591009

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an information embedding method and apparatus for deeply encrypting and embedding into image data auxiliary information related to this image data and to a computer readable recording medium having recorded a program for causing a computer to execute the information embedding method.
2. Description of the Related Art
With advancement in the digitalization of image data, it has become general knowledge that image data, taken from various digital input units such as a digital camera, a scanner and the like, is output with a variety of digital image output units, such as a color printer, a CRT display unit, a photoengraving unit and the like. Furthermore, such image data has been handled in an open environment by readjustment of network infrastructure. With this, the protection of rights, such as a design right and copyright, produced by generation or processing of digital image data has become a problem.
For this reason, a variety of techniques for deeply encrypting and embedding the certified information or copyright information of image data (hereinafter referred to as auxiliary information) into the image data have been proposed (e.g., Japanese Unexamined Patent Publication Nos. 8(1996)-289159, 10(1998)-108180, 9(1997)-214636, etc.). These techniques embed certified information or copyright information into the redundancy portion of image data, whereby only reproducing the image data cannot confirm the embedded information; however, the techniques are capable of reading out the information embedded into the data and performing a display or the like, by employing a device or software for reading out the information. For this deep encryption, the details are described in various references (e.g., Kineo Matsui, “Electronic Watermark,” in O plus E No. 213, August 1997).
As such a technique of deep encryption, various techniques, such as a pixel space utilizing type, a quantization error utilizing type, a frequency area utilizing type and the like, are known. The pixel space utilizing type is a method of taking out, for example, a plane of 3×3 pixels adjacent to an object pixel and embedding auxiliary information into 8 bits around this plane. The quantization error utilizing type is a method of directing attention to a quantization error occurring in the process of compressing image data and of controlling quantized output to an even number and an odd number by the 0 and 1 in the bit sequence of auxiliary information and apparently embedding the auxiliary information into image data as quantization noise. For this quantization error utilizing type, the details are described in “Deep Cryptograph for an Image” (Kineo Matsui, Morikita publication, 1993). The frequency area utilizing type is a method of embedding auxiliary information into a frequency area visually insensitive on an image area. For instance, the high-frequency component in an image is a visually insensitive area, so image data is separated into a plurality of frequency bands, auxiliary information is embedded into a high-frequency band, and furthermore, the frequency bands are reconstructed to the image data, whereby the auxiliary information can be embedded. Also,with regard to human visual property, there is an area that generally color difference or color saturation information is lowered in ability to discriminate gradation more than brightness information and invisible recording becomes possible in the portion of a difference between the brightness information and the color difference or color saturation information. Therefore, auxiliary information can also be embedded into this area.
Besides, a method of embedding auxiliary information into a bit plane whose S/N ratio is poor as an image, by mixing the information with the redundancy of noise; a method of embedding auxiliary information into the redundancy of an information variation in a pixel block (space) in a certain range; a method of embedding auxiliary information into a quantization error when the amount of data information shrinks by coding in performing data compression and the like are enumerated.
Here, a description will be made of an information embedding method that takes advantage of the information conversion redundancy of the pixel block in a density pattern for expressing gradation in a binary image.
There is a density pattern method as a method of expressing multi-level gradation by binary bit mapping.
FIG. 23
is a model case of expressing gradation by a change in the area of 4 binary pixels employed as one unit. Here, the number of gradations, Leq, which can be represented in terms of the number of specified bits, L, and the number of pixels, n×n, can be represented by the following Eq. (1):
L
eq=(
L
−1)
n
2
+1  (1)
Therefore, the case shown in
FIG. 23
can express 5 gradations. Also, as clear from
FIG. 23
, a plurality of patterns indicating a single quantization level (of the same gradation) are present depending upon which of the four pixels puts up a bit that becomes black. For example, quantization levels
0
and
4
can take only a single pattern, but quantization levels
1
and
3
have 4 possible patterns and quantization level
2
has 6 possible patterns. That is, the number of patterns is determined by a combination array of m black pixels and (n
2
−m) white pixels in an array of n
2
pixels, so n
2
Cm possible patterns per gradation are considered. By taking advantage of redundant pixel array patterns with respect to the same quantization level, it is possible to represent another piece of information.
If it is assumed that another piece of information in this case is an integral number, the information will be allocated 4 integral numbers and have 2 bits, for example, in the case of quantization level
1
. That is, in the case of quantization level m, the number of allocated bits, BITm, is calculated as:
BIT
m
=[log
2
(
n
2
Cm
)]  (2)
Therefore, in the case of quantization level
2
,
BIT
2
=[log
2
(2
2
C
2
)]=log
2
6=2.585
from Eq. (2). In addition, the number of quantization bits that can be represented by the entire block of 4 pixels becomes:
BIT
0
+BIT
1
+BIT
2
+BIT
3
+BIT
4
=0+2+2.585+2+0=6.585 bits
In this manner, the pixel position of putting up a bit that becomes black can be represented in the redundancy as a single piece of information. That is, there are only quantization levels
0
and
4
as visual information, but in terms of the redundancy of another piece of information, there are thus various variations in each quantization level.
Note that if the number of pixels constituting each block increases, the redundancy will become great and the amount of information that can be embedded will increase, however, on the other hand, this block will become recognized as a unit beyond the integration effect of visual sense and image quality will be lost. Therefore, the visual property and the amount of information to be embedded are in a trade-off relationship.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a novel method for embedding auxiliary information, such as the above-mentioned, into image data.
If image data in a certain color space is converted to a color space differing in color reproduction characteristic from this color space, distortion will occur in the color space. Because of this distortion, quantization resolution shrinks or increases when mapping of data is performed between both color spaces. For this reason, unless the error is diffused to some degree, a color space cannot be converted. The present invention embeds auxiliary information by taking advantage of the quantization error that occurs when this color space is converted.
That is, an information embedding method according to the present invention is an information embedding method of deeply encrypting and embedding into image data auxiliary information related to said image d

Affiliated with

Also associated with

Rating

Information embedding method, apparatus and recording medium does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Information embedding method, apparatus and recording medium, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information embedding method, apparatus and recording medium will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3076712