Title of Invention

DIGITAL DATA FALSE ALTERATION DETECTION METHOD AND DIGITAL DATA FALST ALTERATION DETECTION APPARATUS

Abstract A digital data false alteration detection program causes a computer to execute (a) a step (S1) of dividing digital data into a plurality of smaller block data, (b) a step (S2) of extracting noise inherent to a digital data acquisition device for each of the small block data, (c) a step (S3) of calculating correlation of the noise between adjacent small block data, and (d) a step (S4) of detecting small block data having noise correlation lower than a level predetermined for the surrounding small block data, as falsely altered data.
Full Text DIGITAL DATA FALSE ALTERATION DETECTION METHOD AND
DIGITAL DATA FALSE ALTERATION DETECTION APPARATUS
TECHNICAL FIELD
The present invention relates to a method and an apparatus for detecting a false
alteration of digital data.
BACKGROUND OF THE INVENTION
An encryption technique called "the digital signature" is available as a method
of detecting a false alteration of digital data. In this technique, the genuineness of digital
data is determined by collating a hash. Generally, the hash is collated by producing a
hash from the current data and comparing the particular hash with the original hash
written in the header portion of the data.
This approach effectively functions in a completely closed system (a system
other than a general-purpose system, using a special data format and having no fixed
users), but cannot meet the requirement of an open system (a general-purpose system
using an ordinary data format and assuming a multiplicity of unspecified users). This is
by reason of the fact that once the file format is changed, the approach becomes
inapplicable any longer.
Another approach called "the electronic watermark" is available. This concerns
a method in which data not related to digital contents are buried in the digital contents
or in which the data buried are extracted and collated. The "electronic watermark"
requires such a structure that the buried data are not easily erased by the process of
editing, compression, transmission or conversion of the contents and that the buried data
cannot be easily falsely altered or overwritten with false information. In this method,

the copyright information or the like, once buried, can be extracted even from falsely
altered data, and therefore the originality of the data can be substantiated. Also, the
falsely altered position can be detected by comparing the falsely altered data with the
substantiated original data.
In this method, however, data are required to be buried in advance, and
therefore a device for burying the data is required. Also, the buried data, which can be
readily extracted as long as a burial method is known, has a low durability. Further, the
data burial unavoidably deteriorates the data quality.
A digital image data acquisition device including an analog-to-digital converter,
on the other hand, has a noise characteristic inherent to the analog-to-digital (A/D)
conversion process.
The output signal of the CCD providing a photoelectric conversion device, for
example, is structurally known to contain noises typically including what is called a
lead-out noise constituting the total noises generated in a CCD element, an analog
circuit of a control system and an A/D converter at time time of reading the charge of the
CCD element and a dark charge noise due to the dark current generated in a well under
the photoelectric surface of the CCD.
Fig. 1 is a diagram for explaining the noises contained in the output of a CCD
element. Fig. 1(A) shows a digital image picked up with a lens cap attached to a digital
camera which is frequency-converted using the two-dimensional FFT (Fast Fourier
Transform), and Fig. 1(B) a solidly black digital image (0 in digital value) produced
using a digital image editing program and frequency-converted by the two-dimensional
FFT. Fig. 1 indicates that the output data of the CCD element contains the noises in
composite fashion.
The digital image data acquisition device including the A/D converter, on the

other hand, has a characteristic inherent to the pixel value of the digital image data in
the A/D conversion process.
JP 2000-36069 A and JP 10-228558 A disclose the technical feature of
registering acoustic signals generated in an automated teller machine (ATM) in its
normal operation as reference data, picking up acoustic signals geerated in the ATM
whileit is working, and detecting filthy mechanical alteration or trouble of the ATM by
comparign acoustic signals picked up with the reference data. JP 10-228558 A further
discloses the technical feature of converting analog data into digital data by means of an
A/D converter (a data acquisition board 56).
Each of the inventions disclosed in JP 2000-36069 A and JP 10-228558 A is
directed to detection of trouble of an apparatus by only comparing acoustic signals
geerated in the apparatus with reference data associated with its normal operation while
the apparatus is working. Namely, in each of the inventions disclosed in JP 2000-36069
A and JP 10-228558 A, characteristics inherent to an acosutic data acquisition device
(especially, an A/D conversion process) are not considered.
On the other hand, the present invention is directed to detection of a false
alteration of digitaldata acquired by a digitaldata acquisition device (especially, an A/D
conversion process) by comparing characteristics of digital data extracted by the digital
data acquisition device with characteristics inherent to the digital data acquisition device.
A comparison between digital data acquired by the digital data acquisition device with
reference data associated with its normal operation is not made.
DISCLOSURE OF THE INVENTION
Accordingly, the problem of the present invention is to provide a program and
an apparatus which works effectively even in an open system by using a characteristic

inherent to the A/D conversion process of a digital data acquisition device and detects a
false alteration of digital data without any device for burying data in advance or
extracting the buried data.
In order to solve the problem described above, according to a first aspect of the
invention, there is provided a digital data false alteration detection program for causing
a computer to detect the false alteration of the digital data acquired by a digital data
acquisition device including a light detector or a sound detector and an A/D converter,
characterized in that the computer is caused to execute:
(a) a step of dividing the digital data into at least two or more small block data;
(b) a step of extracting a noise inherent to the digital data acquisition device for
each of the small block data;
(c) a step of calculating the correlation of the noises between adjacent ones of
the small block data; and
(d) a step of detecting small block data having a noise correlation lower than a
level predetermined for the surrounding small block data, as a falsely altered data.
In the configuration according to the first aspect of the invention, preferably,
the step (b) includes the step of converting each of the small block data into a frequency
domain and extracting the high-frequency component of each small block data as a
noise inherent to the digital data acquisition device, or the step (b) includes the step of
converting each of the small block data into a frequency domain and extracting a
specific high-frequency component of the small block data as a noise inherent to the
digital data acquisition device.
Also, preferably, the step (c) includes the step of calculating an accumulated
value of the noises for each of the small block data and calculating the correlation of the
noises from the difference of the accumulated value of the noises between adjacent ones

of the small block data.
Further, in order to solve the problem described above, according to a second
aspect of the invention, there is provided a digital data false alteration detection
apparatus for causing a programmed computer to detect a false alteration of the digital
data acquired by a digital data acquisition device including a light detector or a sound
detector and an A/D converter, characterized by comprising a data divider for dividing
the digital data into at least two small block data, a noise extraction unit for extracting a
noise inherent to the digital data acquisition device for each of the small block data, and
a false alteration detection unit for calculating the correlation of the noise between
adjacent ones of the small block data and detecting a small block data with the noise
correlation lower than a level predetermined for the surrounding small block data, as
falsely altered data.
In the configuration according to the second aspect of the invention, preferably,
the noise extraction unit converts each of the small block data into a frequency domain
and extracts the high-frequency component of each small block data as a noise inherent
to the digital data acquisition device. As an alternative, the noise extraction unit
converts each of the small block data into a frequency domain and extracts a specific
high-frequency component of each small block data as a noise inherent to the digital
data acquisition device.
Also, preferably, the false alteration detection unit calculates an accumulated
value of the noises for each of the small block data and calculates the correlation of the
noises from the difference of the accumulated value of the noises between adjacent ones
of the small block data.
More preferably, the data divider is adapted to divide the small block data into
data of an arbitrary size. Also, the data divider is adapted to divide the digital data at an

arbitrary position.
Also, in order to solve the problem described above, according to a third aspect
of the invention, there is provided a digital image data false alteration detection program
for causing a computer to detect a false alteration of the digital image data acquired by a
digital image data acquisition device including an A/D converter, characterized in that
the computer is caused to execute:
(a) the step of extracting the noise characteristic of the pixel values of the
digital image data; and
(b) the step of comparing the extracted noise characteristic with the noise
characteristic inherent to the A/D conversion process of the digital image data
acquisition device, and based on the result of comparison, detecting a false alteration of
the digital image data acquired by the digital image data acquisition device.
Also, in order to solve the problem described above, according to a fourth
aspect of the invention, there is provided a digital image data false alteration detection
apparatus for causing a programmed computer to detect a false alteration of the digital
image data acquired by a digital image data acquisition device including an A/D
converter, characterized by comprising an image data noise characteristic extraction unit
for extracting the noise characteristic of the pixel values of the digital image data, and
an image data false alteration detection unit for comparing the noise characteristic
extracted by the image data noise characteristic extraction unit with the noise
characteristic inherent to the A/D conversion process of the digital image data
acquisition device and based on the result of comparison, detecting a false alteration of
the digital image data acquired by the digital image data acquisition device.
Also, in order to solve the problem described above, according to a fifth aspect
of the invention, there is provided a digital image data false alteration detection program

for causing a computer to detect a false alteration of the digital image data acquired by a
digital image data acquisition device including an A/D converter, characterized in that
the computer is caused to execute:
(a) the step of extracting the noise characteristic of the pixel values of the
digital image data; and
(b) the step of dividing the digital image data into at least two small blocks,
comparing the noise characteristics between adjacent ones of the small blocks and upon
development of an anomaly between the compared noise characteristics, detecting a
false alteration of the digital image data acquired by the digital image data acquisition
device.
Also, in order to solve the problem described above, according to a sixth aspect
of the invention, there is provided a digital image data false alteration detection
apparatus for causing a programmed computer to detect a false alteration of the digital
image data acquired by a digital image data acquisition device including an A/D
converter, characterized by comprising a noise characteristic extraction unit for
extracting the noise characteristic of the pixel values of the digital image data, and a
false alteration detection unit for dividing the digital image data into at least two small
blocks, comparing the noise characteristics between adjacent ones of the small blocks
based on the noise characteristic extracted by the noise characteristic extraction unit and
upon development of an anomaly between the compared noise characteristics, detecting
a false alteration of the digital image data acquired by the digital image data acquisition
device.
Also, in order to solve the problem described above, according to a seventh
aspect of the invention, there is provided a digital image data false alteration detection
program for causing a computer to detect a false alteration of the digital image data

acquired by a digital image data acquisition device including at least an A/D converter,
characterized in that the computer is caused to execute:
(a) the step of extracting the characteristic about the pixel values of the digital
image data; and
(b) the step of comparing the extracted characteristic with the characteristic
inherent to the pixel value of the digital image data in the A/D conversion process of the
digital image data acquisition device and based on the result of comparison, detecting a
false alteration of the digital image data acquired by the digital image data acquisition
device.
In the configuration according to the seventh aspect of the invention, preferably,
the step (a) includes the step of extracting a histogram about the pixel values of the
acquired digital image data, and the step (b) includes the step of comparing the
extracted histogram with the inherent histogram about the pixel values of the digital
image data in the A/D conversion process of the digital image data acquisition device,
and in the case where the inherent histogram assumes a continuous value while the
extracted histogram assumes a discontinuous value, detecting a false alteration of the
digital image data acquired by the digital image data acquisition device.
In the configuration according to the seventh aspect of the invention, preferably,
the step (a) includes the step of dividing the acquired digital image data into at least two
equal small blocks and extracting the array pattern of the pixel values of each small
block, and the step (b) includes the step of detecting a false alteration of the digital
image data in the case where the array patterns of the pixel values of the small blocks
extracted in the step (a) are coincident with each other, as compared with the inherent
characteristic that the probability that the array patterns of the pixel values of the small
blocks coincide with each other is very low.

In the configuration according to the seventh aspect of the invention, preferably,
the digital image data acquisition device includes an image acquisition device having a
CCD, and the step (a) includes the step of extracting the pixel value of each pixel of the
acquired digital image data, while the step (b) includes the step of calculating a
predicted pixel value of each pixel of the digital image data by interpolation based on
the CCD matrix array of the digital image data acquisition device from the pixel value
of each pixel of the digital image data extracted in the step (a) and in the case where the
pixel value of each pixel extracted in the step (a) fails to coincide with a corresponding
predicted pixel value, detecting a false alteration of the digital image data.
Also, in order to solve the problem described above, according to an eighth
aspect of the invention, there is provided a digital image data false alteration detection
apparatus for causing a programmed computer to detect a false alteration of the digital
image data acquired by a digital image data acquisition device including at least an A/D
converter, characterized by comprising an image data characteristic extraction unit for
extracting the characteristic about the pixel values of the digital image data, and an
image data false alteration detection unit for comparing the characteristic extracted by
the image data extraction unit with the characteristic inherent to the pixel value of the
digital image data in the A/D conversion process of the digital image data acquisition
device and based on the result of comparison, detecting a false alteration of the acquired
digital image data.
In the configuration according to the eighth aspect of the invention, preferably,
the image data characteristic extraction unit extracts a histogram about the pixel values
of the acquired digital image data, and the image data false alteration detection unit
compares the histogram extracted by the image data characteristic extraction unit with
the histogram inherent to the pixel values of the digital image data in the A/D

conversion process of the digital image data acquisition device, and in the case where
the inherent histogram assumes a continuous value while the histogram extracted by the
image data characteristic extraction unit assumes a discontinuous value, detects a false
alteration of the digital image data acquired by the digital image data acquisition device.
In the configuration according to the eighth aspect of the invention, preferably,
the image data characteristic extraction unit divides the acquired digital image data into
at least two or more equal small blocks and extracts the array pattern of the pixel values
of each small block, and the image data false alteration detection unit detects a false
alteration of the digital image data in the case where the array patterns of the pixel
values of the small blocks extracted by the image data characteristic extraction unit are
coincident with each other, as compared with the inherent characteristic that the
probability that the array patterns of the pixel values of the small blocks coincide with
each other is very low.
In the configuration according to the eighth aspect of the invention, preferably,
the digital image data acquisition device includes an image acquisition device having a
CCD, and the image data characteristic extraction unit extracts the pixel value of each
pixel of the acquired digital image data, while the image data false alteration detection
unit calculates a predicted pixel value of each pixel of the digital image data by
interpolation based on the CCD matrix array of the digital image data acquisition device
from the pixel value of each pixel of the digital image data extracted by the image data
characteristic extraction unit, and in the case where the pixel value of each pixel
extracted by the image data characteristic extraction unit fails to coincide with a
corresponding predicted pixel value, detects a false alteration of the digital image data.
Also, in order to solve the problem described above, according to a ninth
aspect of the invention, there is provided a digital image data false alteration detection

program for causing a computer to detect a false alteration of the digital image data
acquired by a digital image data acquisition device, characterized in that the computer is
caused to execute:
the step of detecting a focused area in an image based on the digital image data,
and upon determination that two or more areas are detected and spaced from each other
by at least a predetermined distance, detecting a false alteration of the digital image data.
Also, in order to solve the problem described above, according to a tenth aspect
of the invention, there is provided a digital image data false alteration detection system
for causing a programmed computer to detect a false alteration of the digital image data
acquired by a digital image data acquisition device, characterized by comprising a
focused area detection unit for detecting focused areas in an image based on the digital
image data, and a false alteration detection unit for detecting a false alteration of the
digital image data upon determination, based on the positions, in the image, of the areas
detected by the focused area detection unit, that a plurality of the areas exist and are
spaced from each other by a predetermined distance.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
Fig. 1 is a diagram for explaining the output noises of a CCD device, in which
(A) shows a digital image picked up by a digital camera with a lens cap, as frequency-
converted using the two-dimensional FFT, and (B) shows a solidly black digital image
(zero in digital value) produced according to a digital image editing program, as
frequency-converted using the two-dimensional FFT.
Fig. 2 is a flowchart of the digital data false alteration detection program
according to a first embodiment in the first aspect of the invention.
Fig. 3 is a diagram showing a state in which a digital image is divided into

small block images.
Fig. 4 is a diagram showing an example of the digital image, in which (A)
shows a falsely altered digital image, and (B) an original digital image not falsely
altered.
Fig. 5 is a graph showing an average accumulated value of the high-frequency
components of the small block image in the digital image versus the variations of the
accumulated value of the high-frequency components of the small block images, in
which (A) corresponds to the digital image of Fig. 4(A) and (B) corresponds to Fig.
4(B).
Fig. 6 is a flowchart of a digital data false alteration detection program
according to another embodiment in the first aspect of the invention.
Fig. 7 is a diagram showing a state in which a digital image is divided into
small block images.
Fig. 8 is a graph showing the spectrum obtained by two-dimensional FFT.
Fig. 9 is a graph showing RMSE for each adjacent points, in which (A) is a
graph corresponding to a falsely altered digital image, and (B) a graph corresponding to
a digital image not falsely altered.
Fig. 10 is a flowchart of a digital data false alteration detection program
according to still another embodiment in the first aspect of the invention.
Fig. 11 is a graph showing the RMSE value for each adjoining point, in which
(A) to (D) correspond to a falsely altered digital image.
Fig. 12 is a graph showing the RMSE value for each adjoining point, in which
(A) and (B) correspond to falsely altered digital data, and (C) corresponds to an original
digital image not falsely altered.
Fig. 13 is a block diagram showing a general configuration of a digital data

false alteration detection apparatus according to an embodiment in the second aspect of
the invention.
Fig. 14 is a graph showing an average accumulated value of the high-frequency
components of the small blocks in the voice data obtained by a computer versus the
variations of the accumulated values of the high-frequency components of the small
blocks, in which (A) corresponds to the falsely altered voice data and (B) corresponds to
the original voice data.
Fig. 15 is a flowchart showing a digital image data false alteration detection
program according to an embodiment in the third aspect of the invention.
Fig. 16(A) is a diagram showing a digital image picked up by a digital camera
and not falsely altered.
Fig. 16(B) is a diagram showing the image of Fig. 16(A) processed for
emphasizing the fluctuation of the least significant bit of the pixel value.
Fig. 17(A) is a diagram showing the image of Fig. 16(A) after false alteration.
Fig. 17(B) is a diagram showing the image of Fig. 17(A) processed for
emphasizing the fluctuation in the same manner as in Fig. 16(A).
Fig. 18 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the fourth aspect
of the invention.
Fig. 19 is a flowchart showing a digital image data false alteration program
according to an embodiment in the fifth aspect of the invention.
Fig. 20 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the sixth aspect
of the invention.
Fig. 21 is a flowchart showing a digital image data false alteration program

according to an embodiment in the seventh aspect of the invention.
Fig. 22 is a flowchart showing a digital image data false alteration program
according to another embodiment in the seventh aspect of the invention.
Fig. 23(A) is a diagram showing a digital image picked up by a digital camera
and not falsely altered.
Fig. 23(B) is a diagram showing a histogram about the pixel values of the
image shown in Fig. 23(A).
Fig. 24(A) is a diagram showing the image of Fig. 23(A) falsely altered by the
gradation conversion process.
Fig. 24(B) is a diagram showing a histogram about the pixel values of the
image of Fig. 24(A).
Fig. 25 is a flowchart showing a digital image data false alteration detection
program according to still another embodiment in the seventh aspect of the invention.
Fig. 26(A) is a diagram showing a digital image picked up by a digital camera
and not falsely altered.
Fig. 26(B) is a diagram showing an image falsely altered by stamping the
image of Fig. 26(A).
Figs. 27(A) to (D) are diagrams showing array patterns of the image portions 1
to 4 in Fig. 26(B) after equally dividing the image of Fig. 26(B) into small blocks and
extracting the array pattern of the pixel values of the small blocks.
Fig. 28 is a flowchart showing a digital image data false alteration detection
program according to yet another embodiment in the seventh aspect of the invention.
Fig. 29 is a diagram showing a CCD matrix.
Fig. 30(A) is a diagram showing an example of the pixel values detected by a
portion of the CCD matrix shown in Fig. 29.

Fig. 30(B) is a list of the CCD devices of a portion of the CCD matrix of Fig.
29 which are numbered by way of explanation.
Fig. 31(A) is a diagram showing the pixel values of a digital image picked up
with a portion of the CCD matrix shown in Fig. 29 and not falsely altered.
Fig. 31(B) is a diagram showing the pixel values of the image of Fig. 31(A)
falsely altered by the Gaussian shading.
Fig. 32(A) is a diagram showing the pixel values assumed to have been
detected by the corresponding CCD devices from the pixel values shown in Fig. 31(B).
Fig. 32(B) is a diagram showing the predicted pixel values calculated from the
pixel values shown in Fig. 32(A).
Fig. 33 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the eighth
aspect of the invention.
Fig. 34 is a flowchart showing a digital image data false alteration detection
program according to an embodiment in the ninth aspect of the invention.
Fig. 35 is a diagram showing a digital image picked up by a digital camera and
not falsely altered.
Fig. 36 is a graph showing the frequency characteristic obtained using HPF by
performing FFT of 16 x 16 pixels extracted from the focused area P4 in the image
shown in Fig. 35.
Fig. 37 is a graph showing the frequency characteristic obtained by the process
similar to that for the area P4 by extracting 16x16 pixels from the unfocused area P5
making up the background of the image shown in Fig. 35.
Fig. 38 is a diagram showing the image of Fig. 35 falsely altered by synthesis
with another image.

Fig. 39 is a graph showing the frequency characteristic obtained by a similar
process to that for the area P4 of Fig. 35 by extracting 16x16 pixels from the area P6
of another image synthesized in the image shown in Fig. 38.
Fig. 40 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according an embodiment in the tenth aspect of
the invention.
BEST MODE FOR CARRYING OUT THE INVENTION
Preferred embodiments of the invention are described below with reference to
the accompanying drawings. Fig. 2 is a flowchart of the digital data false alteration
detection program according to an embodiment in the first aspect of the invention.
According to this embodiment, the digital data is formed of a digital image. With the
digital data false alteration detection program according to this invention, the computer
first divides the digital image of which a false alteration is to be detected, into a
plurality of small block images by a computer, as shown in Fig. 3 (step SI in Fig. 2).
In this example, as apparent from Fig. 3, the digital image is divided into eight small
blocks in horizontal direction and five small blocks in vertical direction for a total of 40
small block images B001 to B040.
Next, each of the small block images divided in step S1 is converted into a
frequency domain using the two-dimensional FFT (Fast Fourier Transform) by the
computer, and the low-frequency component is cut off by a high-pass filter while
amplifying the high-frequency component (step S2 in Fig. 2). The accumulated value of
the high-frequency component is determined for each small block image (step S3 in Fig.
2). Further, the accumulated value of each small block image is compared, and a small
block image having an anomalous value is detected as a falsely altered image (step S4

in Fig. 2). In step S4, preferably, the average of the accumulated value of the small
block images is calculated, and the absolute value of the difference between the
accumulated value and the average value divided by the standard deviation is calculated
for each small block image, so that a block 3 or more off in standard deviation is
detected as a small block image having an anomalous value.
As a specific example, an explanation is given about a case in which a digital
data false alteration detection program according to this invention is installed in the
computer and used for the digital image actually picked up by a digital camera. Fig.
4(B) shows an example of the digital image picked up by a digital camera and not
falsely altered, and Fig. 4(A) the digital image of Fig. 4(B) partially "shaded" according
to the digital image editing program. In Fig. 4(A), reference numeral 10 designates a
portion subjected to the "shading" process. The portion 10 in Fig. 4(A) corresponds to
the small block image B019 in Fig. 3.
Fig. 5 is a graph showing the average accumulated value of the high-frequency
components of the small block images of the digital image versus the variation of the
accumulated value of the high-frequency component of each small block image,
obtained by the computer. (A) corresponds to the digital image of Fig. 4(A), and (B)
corresponds to Fig. 4(B). In Fig. 5, the ordinate represents the absolute value of the
difference between the average accumulated value of the small block imaged and the
accumulated value of each small block image divided by the standard deviation, and the
abscissa the number of each small block. As apparent from Fig. 5, the block B019 of
which the data is falsely altered may be detected as a block of an anomalous
accumulated value of the high-frequency component
In this way, with the digital data false alteration detection program according to
this invention, the computer is caused to detect a false alteration, if any, based on the

noises mixed at the time of acquisition of the digital image. Therefore, a false alteration
of the digital image can be detected without using a device for adding extraneous data
such as an electronic watermark to the digital image of which a false alteration is to be
detected.
Fig. 6 is a flowchart for the digital data false alteration detection program
according to another embodiment in the first aspect of the invention. In this
embodiment, the digital data is formed of a digital image. With the digital data false
alteration detection program according to this aspect of this invention, the computer first
divides the digital image of which a false alteration is to be detected, into a plurality of
small block images as shown in Fig. 7 (step S1 in Fig. 6). In this example, as apparent
from Fig. 7, the digital image is divided into four blocks in horizontal direction and
three blocks in vertical direction for a total of 12 small block images B001 to B012.
Also, the small block images are adjacent to other small block images at the adjoining
points C001 to C017, respectively.
Next, the computer converts each of the small block images into which the
digital image is divided in step S1, into a frequency domain using the two-dimensional
FFT (step S2 in Fig. 6). Fig. 8 is a graph of the spectrum obtained by the two-
dimensional FFT. In Fig. 8, the center peak 20 represents a DC component, and the
higher the frequency, the farther from the peak 20.
The computer smoothes a specific frequency component of all the small block
images B001 to B012 and thus calculates a noise spectrum (step S3 in Fig. 6). In
smoothing a specific frequency component, frequency vectors 23, 24, 25 in a specific
domain are picked out and averaged out along the ordinate 21 in Fig. 8, so that a
spectrum 26 representing the particular domain is obtained. Also, along the abscissa 22,
frequency spectra 27, 28, 29 in a specific domain are picked out and averaged out, so

that a noise spectrum 30 representing the particular domain is obtained.
Further, the computer determines the Euclidean distance (RMSE: Room Mean
Square Error), for each of the adjoining points C001 to C017, of the noise spectrum
between adjoining blocks at the particular adjoining point and detects the block
surrounded by adjoining points having an anomalous RMSE value, as a falsely altered
block (step S4 in Fig. 6).
As a specific example, an explanation is given about a case in which the digital
data false alteration detection program according to the invention is installed in the
computer and used for a digital image actually picked up by a digital camera. Also in
this case, the same digital image as Fig. 4(B) is used as a digital image not falsely
altered. On the other hand, a part of the digital image shown in Fig. 4(B), which is
subjected to "the shading" process by the digital image editing program at a position
corresponding to the block B007 in Fig. 7, is used as a digital image falsely altered.
Fig. 9 is a graph showing the RMSE for each adjoining point determined in
step S4. (A) is a graph corresponding to a digital image falsely altered, and (B) a graph
corresponding a digital image not falsely altered. In Fig. 9, the ordinate represents the
RMSE value, and the abscissa the number of each adjoining point. As apparent from
Fig. 9, the block B007 surrounded by the adjoining points C006, C009, C010, C013
having an anomalous RMSE value can be detected as a block with a falsely altered data.
Also in this embodiment, similar effects to the embodiment shown in Fig. 2 are
obtained.
Fig. 10 is a flowchart for the digital data false alteration detection program
according to another embodiment in the first aspect of the invention. With the digital
data false alteration detection program according to this aspect of the invention, the
computer first divides the digital image of which a false alteration, if any, is to be

detected, into a plurality of small block images as shown in Fig. 7, in the same manner
as in the embodiment shown in Fig. 6 (step S1 in Fig. 10). Next, the computer converts
each of the small blocks divided in step S1 into a frequency domain using the two-
dimensional FFT (step S2 in Fig. 10). As the result of this two-dimensional FFT, as in
the case of the embodiment shown in Fig. 6, a graph of spectrum as shown in Fig. 8 is
obtained.
Then, the computer smoothes a specific frequency component, in the same
manner as in the embodiment shown in Fig. 6, for all the small block images B001 to
B012 thereby to calculate a noise vector (step S3 in Fig. 10). After that, for each
adjoining point, the Euclidean distance (RMSE) of the noise vector between the blocks
adjoining the particular adjoining point, and a block with the RMSE value surrounded
by the anomalous adjoining points is detected as a falsely altered block (step S4 in Fig.
10). In the case where no anomalous adjoining point is detected, the computer changes
both the size of the divided block and the position of dividing the digital image thereby
to move the divided block (step S5 in Fig. 10). Then, steps S1 to S4 are repeated
As a specific example, a case is explained in which the digital data false
alteration detection program according to the invention is installed in the computer and
used actually for the digital image picked up by a digital camera. Also in this example,
the same digital image as in Fig. 4(B) is used as a digital image not falsely altered. A
part of the digital image of Fig. 4(B), which is subjected to the "shading" process by the
digital image editing program at a position corresponding to the block B007 in Fig. 7,
for example, is used as a digital image falsely altered. Figs. 11 and 12 are graphs
showing the RMSE value for each adjoining point determined in step S4. Fig. 11(A) is
a graph in which the "shading" process covers the blocks B002, B003, B006, B007 in
the digital image falsely altered, Fig. 11(B) a graph in which the "shading" process

covers the blocks B003, B004, B007, B008 in the digital image falsely altered, Fig.
11(C) a graph in which the "shading" process covers the blocks B005, B006 in the
digital image falsely altered, and Fig. 11(D) a graph in which the "shading" process
covers the blocks B007, B008 in the digital image falsely altered. Also, Fig. 12(A) is a
graph in which the "shading" process covers only the block B007 in the digital image
falsely altered, Fig. 12(B) a graph in which the "shading" process covers the blocks
B007, B008 in the digital image falsely altered, and Fig. 12(C) a graph corresponding to
the digital image not falsely altered.
As seen from Figs. 11 and 12, an adjoining point having an anomalous RMSE
value is detected in the case of Fig. 12(A), in which a block with the data thereof falsely
altered can be detected by moving the divided blocks.
Although the digital image is converted to a frequency domain by FFT in all
the embodiments described above, the digital image can be converted to a frequency
domain alternatively by the wavelet conversion, DCT (Discrete Cosine Transform) or
DST (Discrete Sine Transform) with equal effect.
Fig. 13 is a block diagram showing a general configuration of a digital data
false alteration detection apparatus according to a first embodiment in the second aspect
of the invention. The apparatus according to this aspect of the invention comprises, as
shown in Fig. 13, a programmed computer, a data divider 40 for dividing the digital
data into a plurality of small block data, a noise extraction unit 41 for extracting a noise
inherent to the digital data acquisition device for each small block data, and a false
alteration detection unit 42 for calculating a noise correlation between adjoining small
block data and detecting small block data having a noise correlation lower than a level
predetermined for the surrounding small block data, as falsely altered data.
The noise extraction unit 41 converts each of the small block data into a
frequency domain and extracts the high-frequency component of each small block data
as a noise inherent to the digital data acquisition device, or converts each of the small
block data into a frequency domain and extracts a specific frequency component of each
small block data as a noise inherent to the digital data acquisition device.
The false alteration detection unit 42 calculates an accumulated value of noises
for each small block data, and thus calculates a noise correlation from the difference of
the accumulated noise value between the adjoining small block data.
The data divider 40 is adapted to divide the small block data into blocks of an
arbitrary size and the digital data at an arbitrary position.
According to still another embodiment in this aspect of the invention, the
digital data false alteration detection program according to the invention is used for the
digital sound data. As compared with the embodiment of Fig. 2 in which the data are
two-dimensionally processed, this embodiment is different only in that the data are
processed one-dimensionally.
In this embodiment, therefore, the digital data false alteration detection
program according to the invention has a flow similar to that of Fig. 2. By use of the
digital data false alteration detection program according to this invention, the computer
first divides the digital sound data of which a false alteration is to be detected, into a
plurality of small blocks. In the example under consideration, the digital sound data
are divided into 132 parts one-dimensionally for a total of 132 small blocks D001 to
D132.
Next, the computer converts each small block division to a frequency domain
using the one-dimensional FFT, while at the same time cutting off the low-frequency
component thereof with a high-pass filter thereby to amplify the high-frequency
component. Then, an accumulated value of the high-frequency component is determined

for each small block. Further, the average accumulated value of each small block is
calculated, and by calculating the absolute value of the difference between the
accumulated value and the average value, divided by the standard deviation, each block
with the standard deviation different by 3 or more is detected as a block having an
anomalous value.
As a specific example, an explanation is given of a case in which the digital
data false alteration detection program according to the invention is installed in the
computer and used for the voice data actually recorded by microphone. The voice data
synthesized by the computer and mixed at a position corresponding to the block D028
of the original voice data recorded is used as a falsely altered voice data.
Fig. 14 is a graph, obtained by the computer, showing the average accumulated
value of the high-frequency component of the small blocks of the voice data versus the
variation of the accumulated value of the high-frequency component of each small
block image. (A) corresponds to the voice data falsely altered, and (B) corresponds to
the original voice data. In Fig. 14, the ordinate represents the absolute value of the
difference between the average accumulated value of the small blocks and the
accumulated value of each small block, divided by the standard deviation, and the
abscissa the number of each small block. As apparent from Fig. 14, the block B028 of
which the data is falsely altered may be detected as a block having an anomalous
accumulated value of the high-frequency component.
Fig. 15 is a flowchart of the digital image data false alteration detection
program according to an embodiment in the third aspect of the invention. As shown in
Fig. 15, with the digital image data false alteration detection program according to the
invention, the computer first extracts the noise characteristic of the pixel value of the
digital image data of which a false alteration is to be detected, acquired by the digital

image data acquisition device including an analog-to-digital converter (A/D converter)
(step S1 in Fig. 15). Next, the computer compares the noise characteristic extracted in
step S1 with the noise characteristic inherent to the A/D conversion process in the
digital image data acquisition device, and based on the result of comparison, detects a
false alteration of the digital image data acquired by the digital image data acquisition
device (step S2 in Fig. 15).
As a specific example, an explanation is given of a case in which the digital
image data alteration detection program according to the invention is installed in the
computer and used for the digital image actually picked up by a digital camera. In this
case, the noise characteristic can be detected by checking the least significant bit of the
pixel value of the digital image data. Nevertheless, it is also possible to extract the noise
characteristic by other appropriate well-known methods. Fig. 16(A) shows a digital
image not falsely altered, picked up by a digital camera, and Fig. 16(B) an image
obtained by processing the digital image shown in Fig. 16(A) for emphasizing the
fluctuation of the least significant bit of the pixel value. According to this embodiment,
in the case where the least significant bit of the pixel value of the original digital image
is 0, the pixel value of the particular pixel is held as it is, while in the case where the
least significant bit is 1, the pixel value of the particular pixel is replaced with 255
thereby to emphasize the fluctuation. Fig. 17(A) shows an image falsely altered from
the image shown in Fig. 16(A) by attaching a copy of a part of the image portion around
the image portion P1 of Fig. 16(A) to a part of the image portion P1, and Fig. 17(B) the
image of Fig. 17(A) processed for emphasizing the fluctuation in a similar manner to
Fig. 16(A).
Due to the noise characteristic inherent to the A/D conversion process of the
digital image data acquisition device, a normal digital image contains noises at random,

and the distribution of the least significant bit of the pixel value is substantially random,
so that the fluctuation appears at random. As shown in Fig. 16(B), therefore, after the
process of emphasizing the fluctuation of the digital image not falsely altered, the image
contains substantially no solidly black area. As understood from Fig. 17(B), however,
once an image is falsely altered, the noise contained in the image portion falsely altered
becomes uniform and the fluctuation is smoothed, thereby the falsely altered image
portion is solidly blackened. In this way, the falsely altered image of Fig. 17(B) is
detected by comparing Fig. 16(B) and Fig. 17(B) with each other.
Fig. 18 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the fourth aspect
of the invention. As shown in Fig. 18, the digital image data false alteration detection
apparatus according to this invention comprises a programmed computer, an image data
noise characteristic extraction unit 50 for extracting the noise characteristic of the pixel
value of the digital image data acquired by the digital image data acquisition device
including an analog-to-digital converter (A/D converter), and an image data false
alteration detection unit 51 for comparing the noise characteristic extracted by the image
data noise characteristic extraction unit 50 with the noise characteristic inherent to the
A/D conversion process of the digital image data acquisition device, and based on the
result of comparison, detecting a false alteration of the digital image data acquired by
the digital image data acquisition device.
Fig. 19 is a flowchart of the digital image data false alteration detection
program according to an embodiment in the fifth aspect of the invention. As shown in
Fig. 19, with the digital image data false alteration detection program according to the
invention, the computer first extracts the noise characteristic of the pixel value of the
digital image data acquired by the digital image data acquisition device (step S1 in Fig.

19). Next, the computer divides the digital image data into at least two small blocks and
compares the noise characteristic between adjoining small blocks, and in the case where
an anomaly develops between the noise characteristics compared, detects a false
alteration of the digital image data acquired by the digital image data acquisition device
(step S2 in Fig. 19).
Fig. 20 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the sixth aspect
of the invention. As shown in Fig. 20, the digital image data false alteration detection
apparatus according to this invention comprises a programmed computer, a noise
characteristic extraction unit 60 for extracting the noise characteristic of the pixel value
of the digital image data acquired by the digital image data acquisition device, and a
false alteration detection unit 61 for dividing the digital image data into at least two
small blocks, comparing the noise characteristics of the adjoining small blocks with
each other based on the noise characteristic extracted by the noise characteristic
extraction unit 60, and in the case where an anomaly appears between the compared
noise characteristics, detecting a false alteration of the digital image data acquired by
the digital image data acquisition device.
Fig. 21 is a flowchart of the digital image data false alteration detection
program according to an embodiment in the seventh aspect of the invention. As shown
in Fig. 21, with the digital image data false alteration detection program according to
the invention, the computer first extracts the noise characteristic of the pixel value of the
digital image data of which a false alteration is to be detected, acquired by the digital
image data acquisition device including an A/D converter (step S1 in Fig. 21). Next, the
computer compares the noise characteristic extracted in step S1 with the noise
characteristic inherent to the pixel value of the digital image data in the A/D conversion

process of the digital image data acquisition device, and based on the result of
comparison, detects a false alteration of the digital image data acquired by the digital
image data acquisition device (step S2 in Fig. 21).
In this case, the characteristic of the pixel value of the digital image data and
the characteristic inherent to the pixel value of the digital image data in the A/D
conversion process of the digital image acquisition device are various.
Fig. 22 is a flowchart of a digital image data false alteration detection program
according to another embodiment in the seventh aspect of the invention. This
embodiment takes into consideration a histogram as a characteristic of the pixel value of
the digital image data and a characteristic of the pixel value of the digital image data in
the A/D conversion process of the digital image data acquisition device. Specifically,
according to this embodiment, as shown in Fig. 22, with the digital image data false
alteration detection program, the computer first extracts a histogram of the pixel value
of the digital image data of which a false alteration is to be detected (step S10 in Fig.
22). Next, the computer compares the extracted histogram with the histogram unique to
the pixel value of the digital image data in the A/D conversion process of the digital
image data acquisition device and in the case where the extracted histogram assumes a
discontinuous value while the inherent histogram assumes a continuous value, detects a
false alteration of the digital image data acquired by the digital image data acquisition
device (step S11 in Fig. 22).
This embodiment is effective especially for detecting a false alteration by the
gradation conversion process of the digital image data.
As a specific example, an explanation is given about a case in which a digital
image data false alteration detection program according to this invention is installed in
the computer and used for me digital image actually picked up by a digital camera. Fig.

23(A) shows an example of the digital image picked up by a digital camera and not
falsely altered, and Fig. 24(A) the digital image of Fig. 23(A) falsely altered by the
gradation conversion of the digital image shown in Fig. 23(A). Fig. 23(B) shows a
histogram of the pixel value of the digital image shown in Fig. 23(A), and Fig. 24(B) a
histogram of the pixel value of the digital image shown in Fig. 24(A). In Figs. 23(B)
and 24(B), the abscissa of the graph represents the gradation value and the ordinate the
frequency.
As shown in Fig. 23(B), the histogram inherent to the pixel value of the digital
image data in the A/D conversion process of the digital image acquisition device
assumes a continuous value. Once the digital image data acquired by the digital image
acquisition device is falsely altered, on the other hand, the histogram of the pixel value
of the particular image data assumes a discontinuous value as shown in Fig. 24(B). In
the case where the histogram assumes a discontinuous value, therefore, a false alteration
of the digital image data can be detected.
This example concerns the case of a digital image in gray scale. As an
alternative, a false alteration of a RGB digital image can also be detected in a manner
similar to the gray scale image by producing a histogram for each channel of R, G, B.
Fig. 25 is a flowchart for a digital image data false alteration detection program
according to still another embodiment in the seventh aspect of the invention.
According to this embodiment, the array pattern of the pixel values for each of the small
blocks into which the digital image data with a false alteration thereof to be detected is
divided is considered as a characteristic of the pixel value of the digital image data, and
compared with the characteristic inherent to the digital image data acquisition device.
Specifically, in this embodiment, with the digital image data false alteration detection
program according to the invention, as shown in Fig. 25, the computer first divides the

digital image data of which a false alteration is to be detected, into at least two or more
small equal blocks, and extracts an array pattern of the pixel values for each small block
(step S20 in Fig. 25). Next, the computer detects a false alteration of the digital image
data in the case where the array patterns of the pixel values of the small blocks extracted
in step S20 coincide with each other in spite of the characteristic fact of the pixel value
of the digital image data in the A/D conversion process of the digital image data
acquisition device to the effect that the probability is very low that the array patterns of
the pixel values of the small blocks coincide with each other (step S21 in Fig. 25).
This embodiment is effective especially for detecting a false alteration of the
digital image data by stamping.
As a specific example, an explanation is given about a case in which the digital
image data false alteration detection program according to the invention is installed in
the computer and used for the digital image actually picked up by a digital camera. Fig.
26(A) shows an example of the digital image picked up by a digital camera and not
falsely altered, and Fig. 26(B) an image falsely altered by stamping from the image
shown in Fig. 26(A). In the stamping process, as shown in the image of Fig. 26(B), an
image portion 1 configured of 3 by 3 pixels is copied and attached to an image portion 2
(corresponding to the image portion P2 in Fig. 26(A)) configured of 3 by 3 pixels.
Similarly, an image portion 3 configured of 3 by 3 pixels is copied and attached to an
image pixel portion 4 (corresponding to the image portion P3 in Fig. 26(A)) configured
of 3 by 3 pixels.
Figs. 27(A) to (D) show array patterns of the pixel values of the image portions
1 to 4 extracted from the equal small blocks of 3 by 3 pixels into which the digital
image is divided. As understood from Fig. 27, the array patterns of the pixel values
coincide between the image portion 1 (Figs. 27(A)) and the image portion 2 (Fig. 27(B)),

and the array patterns of the pixel values coincide between the image portion 3 (Fig.
27(C)) and the image portion 4 (Fig. 27(D)). As a result, a false alteration of the digital
image data can be detected which has been carried out by executing the stamping
process between the image portions 1 and 2 on the one hand, and between the image
portions 3 and 4 on the other hand.
This example represents a case concerning the gray scale digital image.
Nevertheless, a false alteration of a RGB digital image can also be detected in a manner
similar to the gray scale image by extracting the array pattern of the pixel values for
each channel of R, G, B in small blocks.
Fig. 28 is a flowchart for a digital image data false alteration detection program
according to yet another embodiment in the seventh aspect of the invention. This
embodiment is applicable to a digital image data acquisition device having a CCD such
as a digital camera or an image scanner. As shown in Fig. 28, with the digital image
data false alteration detection program according to this invention, the computer first
extracts the pixel value of each pixel of the digital image of which a false alteration is to
be detected, acquired by the digital image acquisition device including an A/D converter
(step S30 in Fig. 28).
Next, the computer calculates a predicted pixel value of each pixel of the
digital image data by the interpolation calculation based on the array of the CCD matrix
of the digital image data acquisition device from the pixel value of each pixel of the
digital image data extracted in step S30, and in the case where the pixel value of a given
pixel extracted in step S30 fails to coincide with a corresponding predicted pixel value,
detects a false alteration of the digital image data (step S31 in Fig. 28). Step S31 is
explained in detail below.
A digital camera, an image scanner, etc. having a CCD, as shown in Fig. 29,

normally comprises a CCD matrix of CCD devices 100 to 102 having R (red), G (green)
and B (blue) filters, respectively, arranged in a predetermined pattern. The CCD devices
100 to 102 of the CCD matrix each correspond to a pixel of the digital image. In this
state, however, each pixel has a pixel value of only one channel of R, G or B (a pixel
corresponding to the CCD device having the R filter, for example, has only a R value),
and therefore, an appropriate digital image cannot be retrieved. In view of this, pixel
values of non-existent channels (the values of G and B for the pixel corresponding to
the CCD device having the R filter, for example) are calculated by interpolation for
each CCD device from the surrounding pixel values of the same channel. Several
methods are available for interpolation calculation. Assuming that the averaging method
is employed in this case, the interpolation calculation is carried out in the following
manner.
Specifically, assume that the interpolation calculation is carried out for the
portion 200 including 4 by 4 CCD devices in the CCD matrix shown in Fig. 29, and that
each CCD device of the portion 200 has detected a pixel value shown in Fig. 30(A).
Incidentally, Fig. 30(B) is a list of the numbers attached to the CCD devices of the
portion 200 to facilitate explanation.
The pixel values of the pixels corresponding to the CCD devices 6, 7, 10 and
11 shown in Fig. 30(B) are determined by interpolation calculation in the following
manner:
(1) Pixels corresponding to the CCD device 6
The R value is determined by averaging the R values of the CCD devices 1, 3,
9 and 11. Specifically, R = (0 + 0 + 0 + 255)/4 = 63.75. The G value, on the other
hand, is determined by averaging the G values of the pixels 2, 5, 7 and 10. Specifically
G = (0 + 0 + 255 + 255)/4 = 127.5. The B value remains as it is, that is, B = 0.

(2) Pixels corresponding to the CCD device 7
The R value is determined by averaging the R values of the CCD devices 3 and
11. Specifically, R = (0 + 255)/2 = 127.5. The G value remains as it is, that is, G = 0.
The B value, on the other hand, is determined by averaging the B values of the CCD
devices 6 and 8. Specifically B = (0 + 255)/2 = 127.5.
(3) Pixels corresponding to the CCD device 10
The R value is determined by averaging the R values of the CCD devices 9 and
11. Specifically, R = (0 + 255)/2 = 127.5. The G value remains as it is, that is, G = 0.
The B value, on the other hand, is determined by averaging the B values of the CCD
devices 6 and 14. Specifically B = (0 + 255)/2 = 127.5.
(4) Pixels corresponding to the CCD device 11
The R value remains as it is, that is, R = 0. The G value is determined by
averaging the G values of the CCD devices 7, 10, 12 and 15. Specifically, (255 + 255 +
0 + 0)/4 = 127.5. The B value, on the other hand, is determined by averaging the B
values of the CCD devices 6, 8, 14 and 16. Specifically, (0 + 255 + 255 + 0)/4 = 127.5.
This interpolation calculation is carried out for all the CCD devices making up
the portion 200 thereby to determine the pixel value of each pixel of a corresponding
digital image as shown in Fig. 31(A).
In this case, the pixel value of the filter of the CCD device is not changed by
the interpolation calculation. As long as the arrangement of the CCD matrix and the
method of interpolation calculation are known in advance, therefore, the pixel value
(predicted pixel value) that a given digital image data should originally hold can be
calculated from the pixel value of the particular digital image data. Unless the digital
image data is falsely altered, the original pixel value and the predicted pixel value of the
digital image data naturally coincide with each other. In the case where the original

pixel value and the predicted pixel value fail to coincide with each other, therefore, a
false alteration of the image data is detected.
This embodiment is effective for detecting a false alteration of the digital
image data committed by any of various well-known false alteration methods.
As a specific example, an explanation is given below about a case in which the
digital image data false alteration detection program according to this invention is
installed in a computer and used for a digital image actually picked up by a digital
camera.
The digital camera is assumed to have the CCD matrix shown in Fig. 29. Fig.
31(B) shows the pixel values extracted from the image falsely altered by the Gaussian
shading process executed on the image shown in Fig. 31(A). Fig. 32(A) shows the pixel
values assumed to have been detected from the pixel values of Fig. 31(B) by the
corresponding CCD devices. Fig. 32(B) shows the predicted pixel values calculated
from the pixel values of Fig. 32(A). The pixel values of Fig. 31(B) fail to coincide with
the predicted pixel values of Fig. 32(B), and therefore a false alteration of the digital
image is detected.
Fig. 33 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the eighth
aspect of the invention. The digital image data false alteration detection apparatus
according to this invention comprises a programmed computer, an image data
characteristic extraction unit 70 for extracting the characteristic of the pixel values of
the digital image data acquired by the digital image acquisition device including an A/D
converter, and an image data false alteration detection unit 71 for comparing the
characteristic extracted by the image data characteristic extraction unit 70 with the
characteristic inherent to the pixel values of the digital image data in the A/D

conversion process of the digital image data acquisition device, and based on the result
of comparison, detecting a false alteration of the digital image data.
According to another embodiment in the eighth aspect of the invention, the
image data characteristic extraction unit 70 extracts a histogram of the pixel values of
the digital image data. The image data false alteration detection unit 71 compares the
histogram extracted by the image data characteristic extraction unit 70 with the
histogram inherent to the pixel values of the digital image data in the A/D conversion
process of the digital image data acquisition device, and in the case where the inherent
histogram assumes a continuous value while the histogram extracted by the image data
characteristic extraction unit 70 assumes a discontinuous value, detects a false alteration
of the digital image data acquired by the digital image data acquisition device.
According to still another embodiment in the eighth aspect of the invention, the
image data characteristic extraction unit 70 divides the digital image data into at least
two or more equal small blocks and extracts the array pattern of the pixel values for
each small block. The image data false alteration detection unit 71, on the other hand,
detects a false alteration of the digital image data in the case where the array patterns of
the pixel values of the small blocks extracted by the image data characteristic extraction
unit 70 coincide with each other, as compared with the unique characteristic that the
probability is very low that the array patterns of the pixel values of the small blocks
coincide with each other.
Yet another embodiment in the eighth aspect of the invention is suitable for
detecting a false alteration of the digital image acquired by a digital image data
acquisition device having a CCD such as a digital camera or an image scanner.
According to this embodiment, the image data characteristic extraction unit 70 extracts
the pixel value of each pixel of the digital image data. The image data false alteration

detection unit 71, on the other hand, calculates a predicted pixel value for each pixel of
the digital image data by the interpolation calculation based on the CCD matrix array of
the digital image data acquisition device from the pixel value of each pixel of the digital
image data extracted by the image data characteristic extraction unit 70, and in the case
where the pixel value of each pixel extracted by the image data characteristic extraction
unit 70 fails to coincide with a corresponding predicted pixel value, a false alteration of
the digital image data is detected.
Fig. 34 is a flowchart for a digital image data false alteration detection program
according to an embodiment in the ninth aspect of the invention. As shown in Fig. 34,
with the digital image data false alteration detection program according to this invention,
a computer detects focused areas in the image based on the digital image data acquired
by the digital image data acquisition device, and upon determination that two or more
detected areas exist and spaced from each other by a predetermined distance, detects a
false alteration of the digital image data (step S1 in Fig. 34).
This embodiment is effective especially for detecting a false alteration of the
digital image committed by the image synthesis process.
As a specific example, an explanation is given below about a case in which the
digital image data false alteration detection program according to this invention is
installed in a computer and used for a digital image actually picked up by a digital
camera. Fig. 35 shows a digital image picked up by a digital camera and not falsely
altered. Fig. 36 is a graph showing the frequency characteristic obtained by extracting
and subjecting to FFT (fast Fourier transform) 16 by 16 pixels in the focused area P4 in
the image shown in Fig. 35 and using a HPF (high-pass filter). Fig. 37 is a graph
showing the frequency characteristic obtained by extracting 16 by 16 pixels in the
unfocused area P5 constituting the background in the image shown in Fig. 35 and

processing them in similar manner to the area P4. Comparison between Fig. 36 and Fig.
37 shows that the spectrum of the high-frequency domain in the focused area is very
strong.
Fig. 38 shows an image falsely altered by the process of synthesizing the image
shown in Fig. 35 with another image. Fig. 39 is a graph showing the frequency
characteristic obtained by extracting 16 by 16 pixels in the area P6 of another image
synthesized with the image shown in Fig. 38 and processing them in a similar manner to
the area P4 shown in Fig. 35. Fig. 39 indicates that the area P6 also has a very strong
spectrum of high-frequency domain and the image is focused. In the image shown in
Fig. 38, therefore, two focused areas (areas P4 and P6) are spaced from each other by a
predetermined distance, thereby indicating that the image is synthesized. Thus, a false
alteration is detected.
Fig. 40 is a block diagram showing a general configuration of a digital image
data false alteration detection apparatus according to an embodiment in the tenth aspect
of the invention. The digital image data false alteration detection apparatus according to
this invention comprises a programmed computer, and as shown in Fig. 40, further
comprises a focused area detection unit 80 for detecting a focused area in an image
based on the digital image data acquired by the digital image data acquisition device
and a false alteration detection unit 81 for detecting a false alteration of the digital
image data in the case where the existence of a plurality of focused areas spaced from
each other by a predetermined distance is determined from the positions of the
particular areas in the image detected by the focused area detection unit 80.
INDUSTRIAL APPLICABILITY
As described above, according to this invention, the presence or absence of a

data alteration is detected by utilizing the characteristic inherent to the analog-to-digital
conversion process of the digital data acquisition device. Even in an open system,
therefore, a false data alteration can be positively detected. Also, the need of a device
for burying the data in advance or extracting the data buried is eliminated. Thus, this
invention greatly contributes to overcoming the problem of false alteration of the digital
data which is expected to be posed by the future development of the IT technology.
WE CLAIM :
1. A method for detection of the false alteration of the digital data acquired by
a digital data acquisition device having a light detector or a sound detector and an A/D
converter, comprising:
(a) a step of dividing said digital data into at least two smaller block data,
(b) a step of extracting noises inherent to said digital data acquisition device for
each of said small block data,
(c) a step of calculating the correlation of said noises between adjacent ones of
said small block data, and
(d) a step of detecting small block data having a noise correlation lower than a
level predetermined for the surrounding small block data, as a falsely altered data.
2. The method as claimed in claim 1, wherein said step (b) comprises a step
of converting each of said small block data into a frequency domain and extracting the
high-frequency component of said each small block data as a noise inherent to said
digital data acquisition device.
3. The method as claimed in claim 1, wherein said step (b) comprises a step
of converting each of the small block data into a frequency domain and extracting a
specific high-frequency component of said each small block data as a noise inherent to
said digital data acquisition device.
4. The method as claimed in any one of claims 1 to 3, wherein said step (c)
comprises the step of calculating an accumulated value of said noise for said each of the

small block data and calculating the correlation of said noise from the difference of the
accumulated value of said noise between adjacent ones of said small block data.
5. A digital data false alteration detection apparatus for causing a
programmed computer to detect a false alteration of the digital data acquired by a digital
data acquisition device including a light detector or a sound detector and an A/D
converter, comprising:
a data divider for dividing said digital data into at least two small block data;
a noise extraction unit for extracting a noise inherent to said digital data
acquisition device for each of said small block data; and
a false alteration detection unit for calculating the correlation of said noise
between adjacent ones of said small block data and detecting the small block data with
said noise correlation lower than a level predetermined for the surrounding small block
data, as a false data.
6. The digital data false alteration detection apparatus as claimed in claim 5,
wherein said noise extraction unit converts each of said small block data into a
frequency domain and extracting the high-frequency component of said each small
block data as a noise inherent to said digital data acquisition device.
7. The digital data false alteration detection apparatus as claimed in claim 5,
wherein said noise extraction unit converts each of said small block data into a
frequency domain and extracting a specific high-frequency component of said each
small block data as a noise inherent to said digital data acquisition device.
8. The digital data false alteration detection apparatus as claimed in any one
of claims 5 to 7, wherein said data divider is adapted to divide said small block data into
data of an arbitrary size.
9. The digital data false alteration detection apparatus as claimed in any one
of claims 5 to 8, wherein said false alteration detection unit calculates an accumulated
value of said noise for each of said small block data and calculating the correlation of
said noise from the difference of the accumulated value of said noise between adjacent
ones of said small block data.
10. The digital data false alteration detection apparatus as claimed in any one
of claims 5 to 9, wherein said data divider is adapted to divide said digital data at an
arbitrary position.
11. A method for detection of a false alteration of the digital image data
acquired by a digital image data acquisition device having an A/D converter,
comprising:
(a) a step of extracting the noise of the pixel value of said digital image data;
and
(b) a step of comparing said extracted noise characteristic with the noise
characteristic inherent to the A/D conversion process of said digital image data
acquisition device, and based on the result of comparison, detecting a false alteration of
the digital image data acquired by said digital image data acquisition device.
12. A digital image data false alteration detection apparatus for causing a
programmed computer to detect a false alteration of the digital image data acquired by a
digital image data acquisition device having an A/D converter, comprising:
an image data noise characteristic extraction unit for extracting the noise
characteristic of the pixel value of said digital image data; and
an image data false alteration detection unit for comparing the noise
characteristic extracted by said image data noise characteristic extraction unit with the
noise characteristic inherent to the A/D conversion process of said digital image data
acquisition device and based on the result of comparison, detecting a false alteration of
the digital image data acquired by said digital image data acquisition device.
13. A method for detection of a false alteration of the digital image data
acquired by said digital image data acquisition device having an A/D converter,
comprising:
(a) a step of extracting the noise characteristic of the pixel value of said digital
image data; and
(b) a step of dividing said digital image data into at least two small blocks,
comparing the noise characteristics between adjacent ones of said small blocks and
upon development of an anomaly between the compared noise characteristics, detecting
a false alteration of the digital image data acquired by said digital image data
acquisition device.
14. A digital image data false alteration detection apparatus for causing a
programmed computer to detect a false alteration of the digital image data acquired by a
digital image data acquisition device having an A/D converter, comprising:
a noise characteristic extraction unit for extracting the noise characteristic of

the pixel value of said digital image data; and
a false alteration detection unit for dividing said digital image data into at least
two small blocks, comparing the noise characteristics between adjacent ones of said
small blocks based on the noise characteristic extracted by said noise characteristic
extraction unit and upon development of an anomaly between the compared noise
characteristics, detecting a false alteration of the digital image data acquired by said
digital image data acquisition device.
15. A method for detection of a false alteration of the digital image data
acquired by a digital image data acquisition device having an A/D converter,
comprising:
(a) a step of extracting the characteristic about the pixel value of said digital
image data; and
(b) a step of comparing said extracted characteristic with the characteristic
inherent to the pixel value of said digital image data in the A/D conversion process of
said digital image data acquisition device and based on the result of comparison,
detecting a false alteration of the digital image data acquired by said digital image data
acquisition device.
16. The digital image data false alteration detection program as claimed in
claim 15, wherein said step (a) comprises a step of extracting a histogram about the
pixel value of said acquired digital image data, and said step (b) comprises a step of
comparing said extracted histogram with the histogram inherent to the pixel value of the
digital image data in the A/D conversion process of said digital image data acquisition
device, and in the case where said inherent histogram assumes a continuous value while
said extracted histogram assumes a discontinuous value, detecting a false alteration of
the digital image data acquired by said digital image data acquisition device.
17. The method as claimed in claim 15, wherein said step (a) comprises a step
of dividing said acquired digital image data into at least two equal small blocks and
extracting an array pattern of the pixel values for said each small block, and said step
(b) comprises a step of detecting a false alteration of said digital image data in the case
where the array patterns of the pixel values of the small blocks extracted in said step (a)
are coincident with each other, as compared with said inherent characteristic that the
probability that the array patterns of the pixel values of said small blocks coincide with
each other is very low.
18. The digital image data false alteration detection program as claimed in
claim 15, wherein said digital image data acquisition device includes an image
acquisition device having a CCD, and said step (a) comprises a step of extracting the
pixel value of each pixel of said acquired digital image data, while said step (b)
comprises a step of calculating a predicted pixel value of each pixel of said digital
image data by the interpolation calculation based on the CCD matrix array of said
digital image data acquisition device from the pixel value of each pixel of the digital
image data extracted in said step (a), and in the case where the pixel value of each pixel
extracted in said step (a) fails to coincide with a corresponding predicted pixel value,
detecting a false alteration of said digital image data.
19. A digital image data false alteration detection apparatus for causing a
programmed computer to detect a false alteration of the digital image data acquired by a

digital image data acquisition device having an A/D converter, comprising:
an image data characteristic extraction unit for extracting the characteristic
about the pixel value of said digital image data; and
an image data false alteration detection unit for comparing the characteristic
extracted by said image data characteristic extraction unit with the characteristic
inherent to the pixel value of the digital image data in the A/D conversion process of
said digital image data acquisition device and based on the result of comparison,
detecting a false alteration of said acquired digital image data.
20. The digital image data false alteration detection apparatus as claimed in
claim 19, wherein said image data characteristic extraction unit extracts a histogram
about the pixel value of said acquired digital image data, and in that said image data
false alteration detection unit compares the histogram extracted by said image data
characteristic extraction unit with the histogram inherent to the pixel value of the digital
image data in the A/D conversion process of said digital image data acquisition device,
and in the case where said inherent histogram assumes a continuous value while said
histogram extracted by said image data characteristic extraction unit assumes a
discontinuous value, detects a false alteration of the digital image data acquired by said
digital image data acquisition device.
21. The digital image data false alteration detection apparatus as claimed in
claim 19, wherein said image data characteristic extraction unit divides said acquired
digital image data into at least two equal small blocks and extracts the array pattern of
the pixel values of said each small block, and in that said image data false alteration
detection unit detects a false alteration of said digital image data in the case where the

array patterns of the pixel values of the small blocks extracted by said image data
characteristic extraction unit are coincident with each other, as compared with said
inherent characteristic that the probability that the array patterns of the pixel values of
said small blocks coincide with each other is very low.
22. The digital image data false alteration detection apparatus as claimed in
claim 19, wherein said digital image data acquisition device includes an image
acquisition device having a CCD, and said image data characteristic extraction unit
extracts the pixel value of each pixel of the acquired digital image data, while said
image data false alteration detection unit calculates a predicted pixel value of each pixel
of said digital image data by the interpolation calculation based on the CCD matrix
array of said digital image data acquisition device from the pixel value of each pixel of
the digital image data extracted by said image data characteristic extraction unit, and in
the case where the pixel value of each pixel extracted by said image data characteristic
extraction unit fails to coincide with a corresponding predicted pixel value, detects a
false alteration of said digital image data.
23. A method for detection of a false alteration of the digital image data
acquired by a digital image data acquisition device, comprising:
a step of detecting focused areas in an image based on said digital image data,
and upon determination that two or more areas are detected and spaced from each other
by at least a predetermined distance, detecting a false alteration of said digital image
data.
24. A digital image data false alteration detection apparatus for causing a
programmed computer to detect a false alteration of the digital image data acquired by a
digital image data acquisition device, comprising:
a focused area detection unit for detecting focused areas in an image based on
said digital image data; and
an alteration detection unit for detecting a false alteration of said digital image
data upon determination, based on the positions, in said image, of said areas detected by
said focused area detection unit, that a plurality of the areas are detected and spaced
from each other by a predetermined distance.
— X —
A digital data false alteration detection program causes a computer to execute
(a) a step (S1) of dividing digital data into a plurality of smaller block data, (b) a step
(S2) of extracting noise inherent to a digital data acquisition device for each of the small
block data, (c) a step (S3) of calculating correlation of the noise between adjacent small
block data, and (d) a step (S4) of detecting small block data having noise correlation
lower than a level predetermined for the surrounding small block data, as falsely altered
data.

Documents:

733-kolnp-2004-abstract.pdf

733-kolnp-2004-assignment.pdf

733-kolnp-2004-claims.pdf

733-KOLNP-2004-CORRESPONDENCE 1.1.pdf

733-KOLNP-2004-CORRESPONDENCE 1.2.pdf

733-KOLNP-2004-CORRESPONDENCE 1.3.pdf

733-kolnp-2004-correspondence.pdf

733-kolnp-2004-description (complete).pdf

733-kolnp-2004-drawings.pdf

733-kolnp-2004-examination report.pdf

733-kolnp-2004-form 1.pdf

733-kolnp-2004-form 18.pdf

733-KOLNP-2004-FORM 27.pdf

733-kolnp-2004-form 3.pdf

733-kolnp-2004-form 5.pdf

733-kolnp-2004-gpa.pdf

733-KOLNP-2004-OTHERS DOCUMENTS 1.1.pdf

733-KOLNP-2004-OTHERS.pdf

733-kolnp-2004-reply to examination report.pdf

733-kolnp-2004-specification.pdf


Patent Number 243335
Indian Patent Application Number 733/KOLNP/2004
PG Journal Number 41/2010
Publication Date 08-Oct-2010
Grant Date 06-Oct-2010
Date of Filing 01-Jun-2004
Name of Patentee FUSO PRECISION CO.LTD
Applicant Address 284, DAIGO-CHO,GOJYO-DORI KARA SUMA NISHIIRU, SHIMOGYO-KU, KYOTO-SHI, KYOTO 600-8106
Inventors:
# Inventor's Name Inventor's Address
1 TAKEDA NAOTO 284, DAIGO-CHO, GOJYO-DORI KARASUMA NISHI-IRU, SHIMOGYO-KY, KYOTO-SHI, KYOTO 600-8106
2 TANIGUCHI KUNIO 566-183, FUJITA, OKAYAMA-SHI, OKAYAMA 701-0221
3 YAMAMOTO KIYOTAKA 1-706,2-1, TAKENOSATO-CHO, OHARANO-HIGASHI, NISHIKYO-KU, KYOTO-SHI, KYOTO 610-1144
4 TANIGUCHI MASAMICHI 566-183, FUJITA, OKAYAMA-SHI, OKAYAMA 701-0221
PCT International Classification Number H04N 1/387
PCT International Application Number PCT/JP2002/126265
PCT International Filing date 2002-12-03
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 2001-368627 2001-12-03 Japan