CN104166957A - Method for embedding and extracting vector graph copyright images - Google Patents
Method for embedding and extracting vector graph copyright images Download PDFInfo
- Publication number
- CN104166957A CN104166957A CN201410258893.XA CN201410258893A CN104166957A CN 104166957 A CN104166957 A CN 104166957A CN 201410258893 A CN201410258893 A CN 201410258893A CN 104166957 A CN104166957 A CN 104166957A
- Authority
- CN
- China
- Prior art keywords
- pel
- images fragment
- embedding
- image
- fragment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The invention relates to a method for embedding and extracting a vector graph copyright images. A copyright image embedding method at present is difficult to resist geometrical attacks of translation, rotation, scaling and combination thereof, and is more difficult to resist interference attacks aimed at primitive addition, modification, deletion and combination thereof. The method for embedding and extracting the vector graph copyright images perform complex proportioning on a vector graph according to a handle value, copyright images are embedded in the complex proportion, resistance of the copyright images to geometrical attacks is guaranteed, a plurality of groups of copyright images are embedded in the vector graph, so that the method in the invention has robustness for interference attacks.
Description
Technical field
The present invention relates to the encrypted copyright resist technology of vector graphics, relate in particular to a kind of embedding and extracting method of vector graphics drm image.The inventive method can be embedded in sightless mode the drm image that represents copyright information in vector graphics file, and can extract in a particular manner drm image, to prove the owner of figure copyright.
Background technology
Vector graphics refers to that the geometric graphic element based on mathematical expression represents computer graphical such as point, line, curve, polygon etc. in utilization.Be different from bitmap, because vector graphics is described with mathematical feature, after its convergent-divergent, there will not be the mosaic feature of bit map type.Vector graphics is because its distortionless zoom feature and accuracy are widely used in the fields such as building, machinery, weaving, electric, building, artistic creation.Its needs that extensively use just, vector graphics is as crystallization of wisdom, creator's painstaking effort, the embodiment of everyone assets especially, the protection problem of its copyright receives much concern always.
The mode of password or digital certificate, due to its not convenient property having, is difficult to extensive popularization.In recent years, be subject to multimedia file to embed the impact of invisible copyright information technology, take the mode that embeds invisible drm image to protect the technology of vector graphics just very more and more to receive people's concern.On the one hand, the drm image of embedding is invisible, and the recipient of file or diffusion person are difficult for discovering; On the other hand; the definitive document that the owner sends to recipient is also unnecessary does extra encryption, and reading figure does not need through decryption step, the disguise of bringing due to this copyright protection technology and the convenience of reading; make the owner and recipient all acceptant, more and more welcome.
But the present situation of drm image embedded technology allows of no optimist, can " be smeared " very easily, be difficult to play the effect of evidence copyright.Existing vector graphics copyright drm image embedded technology, according to the difference of drm image embedding grammar working field, can be divided into transform domain method and Space domain.
In transform domain, mainly by technology such as wavelet transformation, Fourier transforms, realize, this conversion can be resisted the geometric operation such as overall convergent-divergent, translation and rotation, but the attack that can not resist increase, delete and revise.For example, documents 1[health is lush etc. the blind watermarking algorithm based on fractional fourier transform in two-dimensional engineering graphic. and computer utility .2009 (6)] a kind of watermarking algorithm based on fractional fourier transform proposed, the method is extracted the relative coordinate vector of vector graphics middle conductor, structuring one-dimensional complex signal, by fractional fourier transform, obtain the conversion coefficient of complex signal, amplitude adjustment to conversion coefficient is carried out embed watermark, but cannot resist the interference attack such as interpolation, deletion, modification.
In spatial domain, mainly the attributes such as apex coordinate, radius and angle by means of pel embed drm image, according to the pel relying on, mainly be divided into two classes: the first kind, extract the straight line (general containing more straight line in engineering drawing) in engineering drawing, revise the apex coordinate of straight line or straight line is transformed to the embedding that multi-section-line is realized drm image, but the method can not be resisted the attack of convergent-divergent, more cannot resist the attacks such as increase, deletion or modification; Equations of The Second Kind, extract all summits of entity in engineering drawing, modification by means of summit spacing ratio embeds drm image fragment, but cannot resist the attack that increases and delete pel entity, or carry out polar coordinate transform and obtain amplitude and phase place opposing Rotation and Zoom and attack, but can not resist the combination attacks of multiple geometric transformation.For example, documents 2 [Wang Yashun etc. the digital watermarking proportional algorithm based on two-dimensional engineering graphic. University Of Nanchang's journal (industry science version) .2003 (4)] a kind of algorithm that carrys out embed watermark by revising the ratio of each entity summit spacing in engineering drawing proposed, but when increasing or delete pel entity, but watermark sequence accurately cannot be extracted, copyright owner can not be proved.
As above-mentioned, current vector graphics drm image embedded technology is difficult to the geometric attack of opposing to the translation for whole figure, rotation, convergent-divergent, is more difficult to the interference attack that opposing increases or delete pel.Tracing it to its cause, is mainly that these characteristics comprise because the inherent characteristic of vector graphics determines:
(1) the shared byte number of vector graphics file is few;
(2) the metastable transform domain of neither one can be for embedding drm image;
(3) be easy to carry out geometric operation and cause losing drm image fragment;
Therefore, embed drm image more much more difficult than embed drm image in files such as bitmap, video, audio frequency, grid in vector graphics, especially characteristic (3) has been aggravated difficulty that drm image opposing attacks.
In view of this, the present invention proposes a kind of embedding and extracting method of vector graphics copyright drm image, to make up the deficiencies in the prior art.
Summary of the invention
First object of the present invention is: a kind of embedding grammar of vector graphics drm image is provided, and the method can be resisted geometric attack, and interference attack is had to robustness.
Second object of the present invention is: a kind of extracting method of vector graphics drm image is provided, and the method can extract drm image from suffer the vector graphics file of interference attack.
In order to reach above-mentioned purpose, technical scheme of the present invention is as follows:
The embedding of vector graphics drm image and extracting method, comprise embedding step and extraction step, embeds step and comprise:
Step 1.1, structure drm image sequence, will represent the image object discretize of copyright information, and by discrete image data, obtain image sequence; By Segmentation of Image Sequences, it is continuous fragment, be called " images fragment ", the number of images fragment is the length of image sequence, and wish embeds the pel number comprising in vector graphics is the length of this vector graphics, the length that is shorter in length than vector graphics of images fragment;
Step 1.2, the pel comprising for vector graphics builds complex vector operator;
Step 1.3, according to the right complex vector operator of the more adjacent order pel of the generation order of pel (handle value), obtains pel to plural ratio, and this step comparative result is called " front plural ratio ";
Step 1.4, by the part in the front plural ratio of images fragment replacement step 1.3, obtains the new plural ratio of embedded images fragment, is called " middle plural ratio ";
Step 1.5, the middle plural ratio obtaining according to step 1.4 is revised pel pair, and after modification, the right plural ratio of this pel is called " rear plural ratio ", revises result and should make rear plural ratio equal middle plural ratio.
Step 1.6, to step 1.5, is embedded into images fragment the pel centering traversing according to generation order (handle value) traversal step 1.1 of pel one by one;
Extraction step comprises the following steps:
Step 2.1, the pel comprising for vector graphics builds complex vector operator;
Step 2.2, according to the complex vector operator of the more adjacent order pel of the generation order of pel (handle value), obtains pel to plural ratio, is called " extracting plural ratio ", reading images fragment from extract plural ratio;
Step 2.3, to generating order (handle value) traversal step 2.1 to step 2.2, obtains the sample that images fragment forms according to pel;
Step 2.4, analyzing samples, determines correct images fragment and error image fragment;
Step 2.5, is reduced into image by correct images fragment.
For embedding step, preferably, the length of vector graphics and the length ratio of image sequence are more than or equal to 2.
For embedding step, preferably, front plural ratio and middle plural ratio and rear plural ratio are stored as double-precision floating points.
For embedding step, preferably, drm image or images fragment did the encryption process before embedding.
For embedding step, preferably, drm image is replaced the privileged site in front plural ratio.
For embedding step, preferably, revise pel to time, only revise pel centering order pel formerly.
For embedding step, preferably, revise pel to time, only revise pel centering order after pel.
For extraction step, preferably, the traversal direction of extraction is contrary with the traversal direction of embedding.
For extraction step, preferably, the frequency that in analyzing samples, each images fragment occurs, with threshold filtering error image fragment, retains correct images fragment.
For extraction step, preferably, sample and correct images fragment are compared, using the consistent images fragment of comparative result as correct images fragment.
The present invention is embedded into images fragment in the right plural ratio of pel, rather than be directly embedded in the data structure of pel, by revising the right plural ratio of pel, make amended pel equal rear plural ratio to plural ratio, realize the embedding of drm image fragment.Because pel is the ratio between vector operator to plural ratio, the geometric transformation with convergent-divergent, translation, rotation or its combination does not change; Guarantee that the present invention has the beneficial effect that can resist geometric attack.
The generation order (handle value) that the present invention utilizes pel is the foundation of step as a comparison, is the basis with the interference attack that opposing revises, adds, deletes and so on.In the process of creation vector graphics, the handle value that Authoring program distributes according to the time sequencing of pel creation has uniqueness and non-alterability, utilize the generation order of handle value as the traversal reference of images fragment interpolation and extraction, can guarantee the directivity of traversal.While there is interference attack, destruction be the right succession of pel in the certain limit of noise spot present position, do not affect the right order of other pel, make the present invention possess the beneficial effect of certain opposing interference attack.
As mentioned above, adopt after technical scheme of the present invention, the images fragment that is added to vector graphics can be resisted geometric attack completely; Interference attack is had to robustness.Mode below in conjunction with accompanying drawing with embodiment is further elaborated the present invention.
Accompanying drawing explanation
Fig. 1 is that vector figure data is deposited and organized structure schematic diagram;
Fig. 2 is that line segment type pel builds complex vector operator schematic diagram;
Fig. 3 is that multistage line style pel builds complex vector operator schematic diagram;
Fig. 4 is that SPL type pel builds complex vector operator schematic diagram;
Fig. 5 is that circular arc type pel builds complex vector operator schematic diagram;
Fig. 6 is that images fragment is embedded into the schematic diagram in double-precision floating point data structure;
Fig. 7 revises SPL pel schematic diagram;
Fig. 8 revises arc pel schematic diagram;
Fig. 9 is contrast schematic diagram before and after embodiment mono-intends revising with vector graphics;
Figure 10 is contrast schematic diagram before and after embodiment bis-intends revising with vector graphics;
Figure 11 is drm image discretize schematic diagram.
Embodiment
Because the length of byte may be different with hardware condition because of the job platform of computing machine, its length can adjust according to concrete situation, below relate to for example the statement of data length can not be for limiting the present invention's spirit.
As aforementioned, vector graphics refers to that the geometric graphic element based on mathematical expression represents computer graphical such as point, line, curve, polygon etc. in utilization.Its concrete data organizational structure is different because of different plotter programs, but does not depart from the organizational factors shown in Fig. 1.The institutional framework of Fig. 1 signal is that to take the common store structure of dwg form be prototype, select this prototype as the example of elaboration below, can not be with dwg format restriction the present invention's spirit, the embodiment of other vector graphics format outside dwg form, can realize with reference to the present embodiment.
Embodiment mono-
One, embed
Step 1.1
In the present embodiment, the vector graphics that wish embeds drm image comprises
individual pel, its length is
.With the image representation copyright owner after Figure 11 discretize, the image after discretize comprises
individual picture point, all picture points are configured image sequence.In order to prevent that image sequence from illegally being identified, image sequence is done the encryption process, preferred existing cryptographic algorithm (as Arnold conversion), using image sequence as input, obtains encrypted image sequence.In the present embodiment, the length of encrypted image sequence is constant, with encrypt the same long, by
individual picture point forms.Image sequence after encrypting is divided into
individual continuous numerical digit fragment, i.e. images fragment
, acquisition comprises
image sequence
, images fragment
figure place be its length
.For safety, consider images fragment
also can do further encryption, in the present embodiment, omit.The length of compute vectors figure
with image sequence length
ratio, and round downwards
; In the present embodiment,
,
represent can effectively embed in the present embodiment
group image sequence.
Step 1.2
Know and know, the essence of vector graphics is mathematical function image, no matter be line segment, multi-section-line, SPL, circular arc etc. pel, all to draw by necessary parameter the functional picture forming, so angle from function, primitive types can be regarded as type function, and the geometric properties of pel can be regarded as function parameter.
As shown in Figure 1, handle value is given according to the time sequencing of primitive generating by plotter program, arrange in chronological order, there is uniqueness, primitive attribute mark the type (being type function) of the contained pel of this segment data, include but not limited to line segment, multi-section-line, SPL, circular arc, polygon, ellipse, cloud line etc.
In Fig. 1, primitive information has partly been recorded the geometric properties (function parameter) of corresponding pel (type function), for different primitive types, primitive information has partly been recorded the intrinsic call parameter of the type function, for example starting point, terminal, intermediate point, node, the center of circle, start angle, termination point, radius, knot vector etc.
In Fig. 1, out of Memory has partly been recorded the information of the line style, live width, color, figure layer of corresponding pel etc. non-phasor function geometric properties.
According to spirit of the present invention, tackle each pel and build complex vector operator.Now with the pel shown in Fig. 2 to Fig. 5, set forth the structure example of vector operator.
As shown in Figure 2, for line segment ab, the coordinate of getting its starting point a and terminal b builds and characterizes vector
it is as follows,
,
be that a point points to the vector that b is ordered, for the ease of interpreting blueprints, Fig. 2 specially will
shift out.
As shown in Figure 3, for multi-section-line cdefgh, ignore intermediate point and only get its starting point c and terminal h coordinate structure sign vector
it is as follows,
,
to point to from c point the vector that h is ordered.
As shown in Figure 4, for SPL ijklmno, ignore intermediate point and only get starting point i and terminal o structure sign vector
it is as follows,
,
to point to from i point the vector that o is ordered.
As shown in Figure 5, for circular arc pqr, ignore the center of circle and only get initial angle
, end angle
, radius R build to characterize vector
it is as follows,
,
be from p point in the center of circle points to, and phasing degree is
vector.
For circle, can be regarded as initial angle and be
, end angle is
special circular arc, its structure that characterizes vector is with reference to circular arc.
As mentioned above, characterizing vector can use two general formulas below to express
，
，
Formula
be applicable to only get coordinate points and build the situation that characterizes vector, formula
be applicable to extract angle and radius and build the situation that characterizes vector, as for the situation of utilizing other geometric properties (function parameter) to build to characterize vector at this without enumerating, those of ordinary skills can amplify and obtain from above-mentioned publicity.
Above-mentioned formula
and formula
can close and be written as
, by formula
can make its plural form
, formula
be complex vector operator.So, have
, formula
the euler form of complex vector operator, wherein,
Mould is long
, phasing degree
.
At this, the pel of a pair of adjacent order is referred to as " pel to ".
Step 1.3
Spirit according to the present invention, should be according to the right complex vector operator of the more adjacent order pel of the generation order of pel (handle value), can be according to the order comparison from small to large of handle value, and also can be according to handle value order from big to small.That its mode relatively can be taked to do is poor, do the computings such as business, and the present embodiment be take mode that complex vector operator does business and explained as example.The complex vector operator of adjacent order is obtained as business
, the long ratio of mould is
, phase differential is
,
with
be referred to as front plural ratio
so, to getting
or get
participate in follow-up computing and depend on intention; For example, can be in front plural ratio
special result choose, work as
time, choose
as the front plural ratio of adding images fragment
, when
time, choose
as the front plural ratio of adding images fragment
; Also such as, can be in front plural ratio
size choose, work as
time, choose
as the front plural ratio of adding images fragment
, vice versa.The present embodiment will be according to front plural ratio
size choose
or
participate in follow-up computing.
Due to the limitation of computer resource, operation result can only represent with limited figure place, determines the limited of its precision.In order to improve the precision of calculating as far as possible, should select the data type storage calculated factor of high precision.Conventional double-precision floating points is a kind of high-precision data type, as shown in Figure 6, the double-precision floating points of IEEE-754 standard definition, its high 63 is-symbol position S, 62～52 is exponent bits E, low 0～51Wei Shi mantissa position
.According to the present invention's spirit, in the present embodiment, front plural ratio, middle plural ratio and rear plural ratio and extract plural ratio and all will save as double-precision floating point data.
Step 1.4
According to the present invention's spirit, by images fragment
be embedded into front plural ratio
after can obtain new plural ratio, be referred to as middle plural ratio
.Concrete embedded mode has multiple, comprises that the first object of reference decides images fragment
be embedded into
position in the position M of mantissa, as choose primitive types as reference, different primitive types is embedded into different positions; Comprise that the second can adopt the mode of fixed position by images fragment
be embedded into
mantissa position
in.The present embodiment will adopt mode in second to carry out illustrated example.As shown in Figure 6,
mantissa position
in choose
position is arrived
between position
position is interval embeds interval as images fragment
, images fragment embeds interval
position choose excessively do not destroy
precision be advisable with, with embedded images fragment
after vector graphics distortion within reason cause interpreting blueprints error for considering, do not change as far as possible
initial value, guarantee the size numerical value of mark pel constant be bottom line.
Step 1.5
According to the present invention's spirit, should basis
revise pel pair, make amended pel to the rear plural ratio of plural ratio
equal middle plural ratio
; The present invention simultaneously points out preferably only to revise pel centering order pel formerly, and the present embodiment adopts revises pel centering order figure element plan formerly.
As shown in Figure 7, certain pel, to being comprised of pel 20 and pel 30, from handle value, supposes that pel 20 is order pels formerly, existing by certain images fragment
embed in pel 20.As previously mentioned, calculate the ratio of the complex vector operator of pel 30 and pel 20
,
with
.Suppose
, get
for front plural ratio
, use images fragment
replace
image in the position M of mantissa embeds interval
,
.Then, keep pel 20 each points constant, only revise terminal
for
point, will characterize vector
be revised as
, make
with
meet
, complete images fragment
embedding.
As shown in Figure 8, certain pel, to being comprised of pel 40 and pel 50, from handle value, supposes that pel 40 is order pels formerly, existing by certain images fragment
embed in pel 40.As previously mentioned, calculate the ratio of the complex vector operator of pel 50 and pel 40
,
with
.Suppose
, get
for front plural ratio
, use images fragment
replace
image in the position M of mantissa embeds interval
,
.Then, keep center of circle p, radius R, initial angle constant, only revise
point is
point, will characterize vector
be revised as
, make
with
meet
, complete images fragment
embedding.
Step 1.6
According to the spirit of invention, according to the elaboration of earlier figures 7 and Fig. 8 embedded images fragment, as long as travel through all pels pair by the order of handle value, by images fragment sequence
institute comprise each fragment (
) be embedded into one by one in corresponding pel, embedded first last images fragment
after, then from first images fragment
start, the handle position, place of continuing continues to embed, and so circulation, until all
individual pel has all embedded images fragment
till, complete embedding step.
All described in the present embodiment
individual pel is to be general reference, be difficult to by figure expressions of give an example, for show embed before and after the variation of pel, now plan with 12 pels of Fig. 9 count into
, and the change of exaggeration embedding front and back is shown it with image.
In Fig. 9,
the vector graphics of images fragment to be added,
the vector graphics that has added images fragment, vector graphics
all pels all embedded images fragment
, obtain the vector graphics that comprises images fragment
.The pel of label 01～12 sign is respectively: line segment 01, SPL 02, line segment 03, line segment 04, circular arc 05, circle 06, line segment 07, line segment 08, line segment 09, line segment 10, multi-section-line 11, multi-section-line 12.Label 01～12 is combined into vector graphics
, with heavy line, represent; The upper target pel of band " W " is combined into vector graphics
, with thick dashed line, represent;
Two, extract
Step 2.1 is to step 2.3
According to the present invention's spirit, extract images fragment and should be each pel structure complex vector operator.What the present embodiment adopted when adding image is only to revise order pel formerly, instantly (is operating) pel centering order pel formerly
be in back pel to order after pel, to pel centering order pel formerly instantly
embedded images fragment
time change back
in
value, if cause when extracting according to traversal order traversal identical when embedding, what calculate is
, rather than
.So the present embodiment extracts the traversal order of images fragment should contrary with the traversal order of embedded images (starting point of extracting be the terminal embedding), and inverse goes out
.
Take Fig. 9 as example, embedding be last pel of traversing to being pel 11-pel 12, due to multi-section-line 12 when embedding the pel centering order in finally traversing after pel, its value is not affected by embedding operation, extracting traversal ought be from pel 12 and pel
start, first calculate
, obtain
can embed from its image interval
in read last images fragment
.
Extraction traverse second pel to time, should be first by
.First, erase
in image embed interval
, obtain erasing the new plural ratio of images fragment
be referred to as " inverse plural number ratio "; Secondly, revise multi-section-line
, its function parameter being modified (the present embodiment is terminal) is identical with embedding Modification growth function parameter, and after revising, inverse goes out multi-section-line
, and make multi-section-line
meet
, inverse finishes.
Due to
be
in erase image and embed interval
and obtain,
, its corollary is multi-section-line
do not overlap with multi-section-line 11,
.Its error
be referred to as " natural distortion ", natural distortion unavoidably but controlled.Images fragment
length
less, smear rear natural distortion
just less, image embeds interval
the position lower (the closer to lowest order) being positioned, smears rear just less on the impact of initial value; So, control images fragment
the degree of refinement of cutting apart and image embed interval
position, can control nature distortion
size.
Calculate
after, by pel
regard pel 11 as, calculate
.Due to
very little,
even,
, but do not mean that their image embeds interval
unequal.For concrete image, embed interval
position and the degree of refinement cut apart of images fragment,
in the time of in a scope of allowing,
image embed interval
be believable, read and complete the images fragment right to this pel
extraction.
From extraction, traverse the 3rd pel to after starting, with reference to extracting, traverse second operation that pel is right.Suppose instantly to traverse
individual pel and
the pel pair that individual pel forms, a upper extraction pel is to by the
pel and
pel forms, and in upper one extracts operation, calculates
, for instantly extracting operation, only need to be coated with and erase
images fragment embed interval
, obtain inverse plural number ratio instantly
.Revise pel instantly
, obtain amended pel
, and make pel
and pel
meet
, then by pel
regard pel as
, calculate
, from
middle reading images fragment
.
As previously mentioned, extract traverse each pel to time all can be subject to a upper pel pair
impact, along with extracting the carrying out of traversal, each pel is to producing
conduct downwards in capital, makes
constantly amplify.When
while being amplified to a certain degree, so far will have influence on later
image embed interval
confidence level, will
image embed interval
when insincere for the first time, residing traversal is defined as " traversal of breaking one's promise (V) ", by extracting for the first time, traversing all traversal numbers of (comprising end points) before the traversal of breaking one's promise and is referred to as " credible traversal number (NV) ".
The present embodiment vector graphics general reference
individual pel, total completes
inferior embedding, embeds at least
individual complete image sequence
(
round downwards and obtain), in order to guarantee
image embed interval
always credible, image sequence
length
and must meet between credible traversal number (NV)
, often complete once
individual extraction, what the next one extracted is complete image sequence
end images fragment
.Due to
in one group embeds, in end, extract
shi Wuxu inverse can directly extract, and extracting mode is referring to aforementioned last images fragment
extraction.Extract images fragment
until to images fragment
, all must adopt inverse to extract, its extracting mode is referring to aforementioned.
The present embodiment vector graphics general reference
individual pel, the
the end images fragment of group
last images fragment to whole embedding step
between pel number between
between individual, be difficult to determine from extracting traversal starting point, extract and just extract several times the
the end images fragment of group
so, should do
inferior complete extraction traversal; Travel through each time than last time traversal, from traversal starting point to supposition the
organize end images fragment
between equal difference once extract; To guarantee
in individual complete extraction traversal, must there is a traversal to the
the end images fragment of group
definite be correct, this traversal is referred to as " extract just travel through ".In theory, extract just all over obtaining the correct images fragment that number is maximum;
within the specific limits, other extract correct images fragment number that traversal obtains be not necessarily less than extraction just traveling through the correct images fragment number arriving.
the sum of the images fragment that inferior traversal obtains forms sample.
Step 2.4
Analyzing samples by each images fragment and correct image sequence comparison, can match in image sequence, is correct images fragment, is retained; What can not match will not retain.The images fragment frequency that statistics remains into, is zero if there is the frequency of images fragment, illustrates that the interference attack suffering is overweight, extracts unsuccessfully, if the frequency of all images fragments is all non-vanishing, can correctly restore drm image.
Step 2.5
Each images fragment string is good, by decryption processing, restore the drm image after discretize.
Embodiment bis-
One, embed
The present embodiment is than embodiment mono-, and step 1.1～step 1.4 is identical; From step 1.5, different, embodiment mono-revises pel centering order pel formerly, and the present embodiment is to revise pel centering, order after pel, and difference thus, the extraction step of the present embodiment and have difference.When the present embodiment embeds traversal, also according to image sequence
embed one by one, embedded one group of image ordered series of numbers
rear continuation embeds traversal point and continues to embed new one group of image sequence
, with embodiment mono-, after the present embodiment has embedded, in whole vector graphics, comprise
individual complete image sequence
.
This enforcement plan is with 12 pels of Figure 10 to exaggerate it, and signal pel is in the change that embeds front and back.The present embodiment is revised be pel centering order after pel, than Fig. 9 of embodiment mono-, the present embodiment has lacked
, many
; Two, extract
Because the present embodiment selected to revise pel centering order after pel, compare with embodiment mono-, the present embodiment completes after embedding, directly calculate the right plural ratio of pel and can extract images fragment, without inverse, extract, the order identical with embedding traversal order (the starting point of extracting traversal is the starting point that embeds traversal) of extracting traversal, and unnecessary doing
inferior traversal.
Obtain after the sample of this enforcement, generally only need to add up the frequency that each images fragment occurs, by threshold filtering, fall wrong images fragment, retain correct images fragment group string.If know clearly encryption while embedding, after deciphering, restore the drm image after discretize, if do the encryption process, directly do not obtain the drm image after discretize.
Embodiment tri-
The present embodiment does not lie in to exemplify and embeds and extract, and how the present embodiment can resist geometric attack if mainly being set forth the present invention, why interference attack is had to robustness.
One, the opposing of geometric attack
Geometric attack refers to does convergent-divergent of overall importance, rotation, translation or combinations thereof operation to whole vector graphics, with intention destruction after revising pel function parameter, is embedded into the images fragment in function parameter; After convergent-divergent of overall importance, rotation, translation or its combination operation, although pel function parameter changes, figure integral body is constant, does not impact, therefore the method is first choice of assailant to reading.
The present invention builds complex vector operator, calculate pel to plural ratio, and images fragment is embedded in rear plural ratio, as long as the mould of the sign vector of each pel is long and phasing degree is constant, the images fragment in plural ratio is constant thereafter.The operation of convergent-divergent, rotation, translation or its combination, all belongs to linear transformation, the right plural constant rate of same pel before and after conversion, and rear plural constant rate before and after converting, what images fragment was not converted affects, and can resist geometric attack completely.
Previous embodiment one and embodiment bis-, as setting forth exemplifying of the present invention, can resist geometric attack completely.
Two, the robustness of interference attack
Interference attack refers to adding, delete or revising pel, and combination operation; To destroy the mode of pel, the calculating based on pel is cut off, and cause, extract confusion, disturb and extract result.Implement interference attack person, also can consider the destruction disturbing former figure, for example, add pel, should be taken into account whether the pel being added to can affect reading, if impact is read, cause the value of figure to be difficult to embody, the meaning of its attack is also all gone.Equally, modification or deletion pel also should be considered attack result and whether can cause the conversion of figure and depart from original design idea, if cause the loss of original design idea, not only lose the meaning of attack, more may change the disastrous effect causing by Design Imported.As can be seen here; even if the excessive images fragment that causes of interference attack intensity cannot extract, also often because of attack itself, cause figure to be worth without depositing, when figure is worth without depositing or even introduce after purposes risk; the owner for figure, its copyright protection also loses meaning.Thus, significant interference attack is only the meaning of robustness of the present invention, and those interference attack unintentionally, even if the present invention does not possess robustness to its attack, also do not mean that the protection of copyright will under fire affect.
Suppose embodiment mono-to implement to add to attack.Because being later than embodiment mono-pel, the pel being added to generates, after its handle value comes embodiment mono-pel.Embodiment mono-does when extracting
inferior complete extraction traversal, has added how many pels no matter add to attack,
in inferior complete extraction traversal, extract and just travel through and certainly exist, as for extracting just traveling through, be
which in inferior complete extraction traversal, relevant with the pel number adding.According to the method for sample analysis of embodiment mono-, must correctly restore the drm image after discretize, add to attack originally and keep out completely.
Suppose that mono-implementation modification is attacked to embodiment.Owing to having at least
organize complete image ordered series of numbers, each complete image sequence is extracting
time do not rely on previous image sequence
inverse, and there is independence between group.Revise to attack in that group that only can cause revising pel loca after pel lose inverse basis, manufacture wrong images fragment, the pel of revising outside the group of pel place is unaffected.
as long as there is one group not suffered to revise attack in group,
in inferior traversal, must once just extract and travel through, according to the method for sample analysis of embodiment mono-, must correctly restore the drm image after discretize.If revise, attack practical range wide, in other words
large not, can not resist to revise and attack, therefore embodiment mono-can resist modification attack, to revising to attack, there is robustness under certain condition.
Suppose embodiment mono-to implement to delete to attack.The same with delete attacking, delete attack in that group that only can cause deleting pel loca after pel lose inverse basis, manufacture wrong images fragment, the pel outside the group of deletion pel place is unaffected.
as long as there is one group not suffered to delete attack in group,
in inferior traversal, must once just extract and travel through, according to the method for sample analysis of embodiment mono-, must correctly restore the drm image after discretize.If delete, attack practical range wide, in other words
large not, can not resist to delete and attack, therefore embodiment mono-can resist deletion attack, to deleting to attack, there is robustness under certain condition.
Suppose embodiment bis-to implement to add to attack.Because embodiment bis-does not need can directly to calculate pel through inverse in leaching process, plural ratio is not obtained to images fragment, after the pel of attacking again will be arranged in by handle, add due to interpolation
individual pel only produces
the images fragment of individual mistake.According to the method for sample analysis of embodiment bis-, as long as threshold value is greater than
, add to attack the error image fragment producing filtrations that will be done, interpolation attack can be resisted completely, must correctly restore the drm image after discretize.When interpolation is attacked too acutely, the pel number being added to even exceeds the length of former vector graphics
, with the method for sample analysis of embodiment bis-, will lose efficacy; Now, use the matching process in embodiment mono-sample analysis instead, can also correctly restore the drm image after discretize, add to attack and resisted completely.
Suppose that bis-implementation modifications are attacked to embodiment.Because embodiment bis-does not need can directly to calculate pel through inverse in leaching process, plural ratio is not obtained to images fragment, revise to attack and only can cause modifying point place pel to occurring wrong images fragment, as for unaffected outside modifying point.Owing to revise attacking, be to reduce the make the mistake increase of images fragment number of correct images number of chips, when modification is attacked too acutely, the frequency of correct images fragment may not reach the reserve of threshold value, the drm image failure that causes reducing after discretize.Suppose to revise to attack and changed embodiment bis-each group image sequences
first pel pair, will cause losing the basis of extracting computing; But,
as long as have one group not suffered to revise attack in group, according to the method for sample analysis of embodiment mono-, must match correct images fragment, the drm image successfully reducing after discretize.Embodiment bis-can only resist under certain condition to revise and attack, and to revising to attack, has robustness.
Suppose embodiment bis-to implement to delete to attack.Delete to attack only can cause and delete some place pel to occurring wrong images fragment, unaffected as for deleting outside point.Owing to delete attacking, be to reduce the make the mistake increase of images fragment number of correct images number of chips, when deletion is attacked too acutely, the frequency of correct images fragment may not reach the reserve of threshold value, the drm image failure that causes reducing after discretize.Suppose to delete to attack and changed embodiment bis-each group image sequences
first pel pair, will cause losing the basis of extracting computing; But,
as long as have one group not suffered to delete attack in group, according to the method for sample analysis of embodiment mono-, must match correct images fragment, the drm image successfully reducing after discretize.Embodiment bis-can only resist under certain condition to delete and attack, and to deleting to attack, has robustness.
Suppose the combination attacks of adding, revising, delete implementing, as previously mentioned, bis-pairs of independent interference attack of embodiment mono-and embodiment all have robustness, for combination attacks, as long as
in group, there is one group not attacked, during sample analysis, must match complete image sequence, to the intrinsic robustness of combination attacks.
By the elaboration of above-described embodiment, the present invention is clear, and open fully full and accurate, those of ordinary skills implement the present invention and obtain producing effect again of the present invention.Any those of ordinary skill in the art, the apparent modification that the present invention is made all should be regarded the category that does not depart from the present invention as, and protection scope of the present invention is as the criterion with claims.
Claims (10)
1. the embedding of vector graphics drm image and extracting method, is characterized in that comprising embedding step and extraction step, embeds step and comprise:
Step 1.1, structure drm image sequence, will represent the image object discretize of copyright information, and by discrete image data, obtain image sequence; By Segmentation of Image Sequences, it is continuous fragment, be called " images fragment ", the number of images fragment is the length of image sequence, and wish embeds the pel number comprising in vector graphics is the length of this vector graphics, the length that is shorter in length than vector graphics of images fragment;
Step 1.2, the pel comprising for vector graphics builds complex vector operator;
Step 1.3, according to the right complex vector operator of the more adjacent order pel of the generation order of pel (handle value), obtains pel to plural ratio, and this step comparative result is called " front plural ratio ";
Step 1.4, by the part in the front plural ratio of images fragment replacement step 1.3, obtains the new plural ratio of embedded images fragment, is called " middle plural ratio ";
Step 1.5, the middle plural ratio obtaining according to step 1.4 is revised pel pair, and after modification, the right plural ratio of this pel is called " rear plural ratio ", revises result and should make rear plural ratio equal middle plural ratio;
Step 1.6, to step 1.5, is embedded into images fragment the pel centering traversing according to generation order (handle value) traversal step 1.1 of pel one by one;
Extraction step comprises the following steps:
Step 2.1, the pel comprising for vector graphics builds complex vector operator;
Step 2.2, according to the complex vector operator of the more adjacent order pel of the generation order of pel (handle value), obtains pel to plural ratio, is called " extracting plural ratio ", reading images fragment from extract plural ratio;
Step 2.3, to generating order (handle value) traversal step 2.1 to step 2.2, obtains the sample that images fragment forms according to pel;
Step 2.4, analyzing samples, determines correct images fragment and error image fragment;
Step 2.5, is reduced into image by correct images fragment.
2. the method for claim 1, is characterized in that embedding in step drm image or images fragment and did the encryption process before embedding.
3. the method for claim 1, is characterized in that embedding drm image in step and replaces the privileged site in front plural ratio.
4. the method for claim 1, it is characterized in that embedding in step, revise pel to time, only revise pel centering order pel formerly.
5. the method for claim 1, it is characterized in that embedding in step, revise pel to time, only revise pel centering order after pel.
6. the method for claim 1 to 5 any one, is characterized in that embedding the length of vector graphics and the length ratio of image sequence in step and is more than or equal to 2.
7. the method for claim 6, is characterized in that embedding in step front plural ratio and middle plural ratio and rear plural ratio and is stored as double-precision floating points.
8. the method for claim 1, is characterized in that the traversal direction of the traversal direction extracted in extraction step and embedding is contrary.
9. claim 1 or 8 method, is characterized in that the frequency that in extraction step, in analyzing samples, each images fragment occurs, with threshold filtering error image fragment, retains correct images fragment.
10. claim 1 or 8 method, is characterized in that in extraction step sample and correct images fragment to compare, using the consistent images fragment of comparative result as correct images fragment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410258893.XA CN104166957B (en) | 2014-06-12 | 2014-06-12 | Method for embedding and extracting vector graph copyright images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410258893.XA CN104166957B (en) | 2014-06-12 | 2014-06-12 | Method for embedding and extracting vector graph copyright images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104166957A true CN104166957A (en) | 2014-11-26 |
| CN104166957B CN104166957B (en) | 2017-02-15 |
Family
ID=51910750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410258893.XA Expired - Fee Related CN104166957B (en) | 2014-06-12 | 2014-06-12 | Method for embedding and extracting vector graph copyright images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104166957B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106056528A (en) * | 2016-05-31 | 2016-10-26 | 华中科技大学 | Engineering drawing copyright protection and modification positioning method |
| CN106384024A (en) * | 2016-08-31 | 2017-02-08 | 南宁市勘察测绘地理信息院 | Copyright protection method of DWG format data and production method of DWG format data |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050105760A1 (en) * | 2001-12-14 | 2005-05-19 | Eggers Joachim J. | Data embedding and extraction |
| CN103366335A (en) * | 2013-07-17 | 2013-10-23 | 兰州交通大学 | Vector point spatial data full-blind watermarking method based on grid dividing |
-
2014
- 2014-06-12 CN CN201410258893.XA patent/CN104166957B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050105760A1 (en) * | 2001-12-14 | 2005-05-19 | Eggers Joachim J. | Data embedding and extraction |
| CN103366335A (en) * | 2013-07-17 | 2013-10-23 | 兰州交通大学 | Vector point spatial data full-blind watermarking method based on grid dividing |
Non-Patent Citations (2)
| Title |
|---|
| 康菁菁等: "二维工程图中基于分数傅里叶变换的盲水印算法", 《计算机应用》 * |
| 汪亚顺等: "基于二维工程图的数字水印比例算法", 《南昌大学学报（工科版）》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106056528A (en) * | 2016-05-31 | 2016-10-26 | 华中科技大学 | Engineering drawing copyright protection and modification positioning method |
| CN106384024A (en) * | 2016-08-31 | 2017-02-08 | 南宁市勘察测绘地理信息院 | Copyright protection method of DWG format data and production method of DWG format data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104166957B (en) | 2017-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108961141B (en) | Vector map double zero watermarking method, system, storage medium and server | |
| Wang et al. | A high capacity reversible data hiding method for 2D vector maps based on virtual coordinates | |
| Peng et al. | Multipurpose watermarking for vector map protection and authentication | |
| Chen et al. | Comparison of the Bergman and Szegö kernels | |
| CN116385250B (en) | Track data double watermarking method based on robust watermarking and fragile watermarking | |
| Cao et al. | A recursive embedding algorithm towards lossless 2D vector map watermarking | |
| CN103377457B (en) | A kind of vector geographic data Hard Authentication vulnerable watermark method | |
| CN103325082A (en) | Vector map reversible information hiding method based on LSD planes | |
| CN110599478B (en) | Image area copying and pasting tampering detection method | |
| CN104166957A (en) | Method for embedding and extracting vector graph copyright images | |
| CN116805069B (en) | Track data zero watermark generation method, track data zero watermark detection device and storage medium | |
| Wang et al. | RST invariant fragile watermarking for 2D vector map authentication | |
| CN108470318B (en) | Three-dimensional grid double watermarking method based on grouping strategy and neighborhood relation positioning | |
| CN104166956A (en) | Method for embedding and extracting vector graph copyright characters | |
| CN103377320A (en) | Method for selective certification of vector geographic data semi-fragile watermarking | |
| Ren et al. | A zero-watermarking scheme based on spatial topological relations for vector dataset | |
| CN101887575A (en) | Multi-geometric attack resistant robust watermarking method | |
| CN104103031B (en) | Based on normalized Vector spatial data blind watermark method | |
| CN111667394B (en) | Map scaling inference method based on feature description | |
| CN102682239A (en) | Digital watermark technology based copyright protection method orienting to remote sensing tile data | |
| CN102622429B (en) | Geographic database watermarking method based on characteristic distance ratio | |
| Sun et al. | Lossless digital watermarking scheme for image maps | |
| Lafaye et al. | Invisible graffiti on your buildings: Blind and squaring-proof watermarking of geographical databases | |
| Tan | A 3D Model Asymmetric Watermarking Algorithm Based on Optimization Statistics | |
| CN105550970B (en) | A kind of map vector reversible information hidden method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170215 Termination date: 20180612 |