CN104504745B - A kind of certificate photo generation method split based on image and scratch figure - Google Patents
A kind of certificate photo generation method split based on image and scratch figure Download PDFInfo
- Publication number
- CN104504745B CN104504745B CN201510022078.8A CN201510022078A CN104504745B CN 104504745 B CN104504745 B CN 104504745B CN 201510022078 A CN201510022078 A CN 201510022078A CN 104504745 B CN104504745 B CN 104504745B
- Authority
- CN
- China
- Prior art keywords
- image
- background
- split
- certificate photo
- prospect
- 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.)
- Active
Links
Abstract
The present invention relates to technical field of image processing, and the invention discloses a kind of certificate photo generation methods split based on image and scratch figure, specifically include following step：The first step, the foreground and background for marking image, obtain drawing of seeds picture；The portrait that wherein prospect is face, hair and clothes for upper half of body form, background are not cover other regions of portrait；Second step, by drawing of seeds picture（）Three passages be quantized into histogram；3rd step, structure Image Segmentation Model are split the histogram that second step obtains, foreground picture and Background after being split；4th step, the foreground picture for obtaining the 3rd step are mixed with the background colour specified, and obtain corresponding certificate photo image.This method can with fast and high quality generate certificate photo image, can be used for any image capture device, be especially suitable for the application of mobile phone photograph.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of certificate photo generations split based on image and scratch figure
Method.
Background technology
Usual people, which shoot certificate photo image, to be needed to specified spot for photography, is taken time and effort.This is because certificate photo pair
Image request is quite high, and conventional method or camera are shot the picture come and cannot be met the requirements.However appointed place must be arrived
The mode of shooting obviously can not meet the needs of user.
The content of the invention
It is needed for certificate photo acquisition methods of the prior art to specified spot for photography, the technology taken time and effort is asked
Topic, the invention discloses a kind of certificate photo generation methods split based on image and scratch figure.
The goal of the invention of the present invention is realized by following technical proposals：
A kind of certificate photo generation method split based on image and scratch figure, specifically includes following step：The first step, mark
Remember the foreground and background for image, obtain drawing of seeds picture；The people that wherein prospect is face, hair and clothes for upper half of body form
Picture, background are not cover other regions of portrait；Three passages of drawing of seeds picture (R, G, B) are quantized into histogram by second step；
3rd step, structure Image Segmentation Model are split the histogram that second step obtains, foreground picture and background after being split
Figure；4th step, the foreground picture for obtaining the 3rd step are mixed with the background colour specified, and obtain corresponding certificate photo image.It should
Method can with fast and high quality generate certificate photo image, can be used for any image capture device, be especially suitable for mobile phone photograph
Using.
Further, the above method further includes：When the image more than one prospect or background after segmentation, will not connect
Logical region merges, until containing only a prospect and a background.The characteristics of certificate photo is only to include a prospect (portrait),
With a background.
Further, the above method further includes：When there are during zone of ignorance, zone of ignorance is carried out at stingy figure after merging
Reason.To obtained prospect, background image generates a ternary diagram, includes prospect, background, zone of ignorance.A kind of region having more
For zone of ignorance, that is, region to be scratched, certificate photo image can be generated in conjunction with coherent matting technique.
Further, above-mentioned stingy figure is as follows：Firstth, image and corresponding ternary diagram are obtained；Secondth, count
Image gradient is calculated, image is sampled according to image gradient, each point obtains one group of prospect, background candidate to zone of ignorance
Point set；3rd, a pair of of prospect, background dot are selected from candidate point set, calculates a four-tuple.
Further, the above method, which is further included, is extended ternary diagram according to the distribution of color of image, further contracts
Small zone of ignorance.
Further, the above method further includes prospect, the background dot that optimization sampling obtains, and calculates a new quaternary
Group.
Further, the above method is further included is smoothed to calculating four-tuple.
Further, the above method is further included using the optimal foreground image of fast algorithm approximate calculation.
Further, image is split using maximum-flow algorithm in above-mentioned image segmentation.
Further, the above-mentioned mode being smoothed to four-tuple is to be smoothed using Gaussian function.
By using above technical solution, the beneficial effects of the invention are as follows：The present invention by by mobile phone or other photograph
The portrait image that camera is shot under arbitrary background environment is first quantified as histogram, then carries out image segmentation, obtains portrait prospect
It is mixed afterwards with the background color of needs (such as red, blueness or white), so as to quickly and easily obtain certificate photo
Image.This method can with fast and high quality generate certificate photo image, can be used for any image capture device, be especially suitable for hand
Machine is taken pictures application.
Description of the drawings
Fig. 1 is the flow chart of the certificate photo generation method split based on image and scratch figure of the present invention.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, below in conjunction with specific embodiment, to this
Invention is described in more detail.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Fig. 1 is the flow chart of the certificate photo generation method split based on image and scratch figure of the present invention.
One of embodiment
It based on image splits and scratches the certificate photo generation method of figure, specifically comprises the following steps：
Step 1: marking the foreground and background of image, drawing of seeds picture is obtained；Wherein prospect for face, hair and on
The portrait of half body clothes composition, background are not cover other regions of portrait.The method of the wherein mark of prospect is：Pass through face
Key point fixation and recognition goes out the key point of face in image, and then left and right connects two and extends to some hairs of covering outward,
It is vertically connected with, until clothes for upper half of body since the crown covers some hairs.The mark of background then selects not including covering people
Any other region of picture.In the case where combining the technology of face key point location, these markers works can be completed automatically and accurately,
User is not required to participate in mark, therefore can accomplish to be completely automatically generated certificate photo image.(identify in the picture face, with
And key point, such as hair, eyes clothes etc. belong to the prior art in face, are not belonging to the emphasis of the present invention, herein not
It is described in detail, similarly clothes for upper half of body is also in this way, can be realized by the methods of setting of threshold value).
Step 2: three passages of drawing of seeds picture (R, G, B) are quantized into histogram.Show that drawing of seeds picture institute is actual simultaneously
The number of the Bin (binary system) used, the number of the Bin of actual use cut the structure on model side for figure in step 4.
Such as：To each pixel I in every drawing of seeds pictureI=1,2 ..., H*W(R, G, B) is in accordance with the following methods by its amount
Change into 64 × 64 × 64=262144 Bin, initialize a two-dimensional array Zx,y, for recording the quantization of corresponding pixel points
Index, wherein：H, W represent the height and width of drawing of seeds picture respectively.Initialize histogram：
Calculating space coordinates be x, the quantization index of the pixel value R, G, B of y：
Index=B/256.0*64+64*G/256.0*64+4096*R/256.0*64 (2)
Statistic quantification result：
H262144[index]=H262144[index]+1 (3)
Zx,y=H262144[index] (4)
The number used of the histogram Bin of actual use is finally calculated, computation rule is statistic histogram H262144
It is not 0 quantity in [index].
Step 3: calculate the global variance of drawing of seeds picture：
In formula, IiRepresent ith pixel, ΩiThe neighborhood of i is designated as under expression pixel, NUM is representedActual meter
Calculate number.(calculating of global variance is mainly used for the calculating that figure in step 4 cuts model side in this step).
Step 4: structure Image Segmentation Model is split image, foreground picture and Background after being split.
(1) color of image and space length are calculated, the side of model is cut according to color of image and space length structure figure：
Calculate the color distance of current pixel point i and neighborhood territory pixel point j
Calculate the space length of current pixel point i and neighborhood territory pixel point j
Calculating figure cuts the weight on model side：
By image current pixel point i (0..H*W-1) and neighborhood territory pixel point j ∈ ΩiA line is connected, the weight on side isWherein alpha meets more than or equal to 0.8 and is 1 less than the sum of 1, alpha and beta.
(2) amount of the number used of the histogram Bin of the actual use calculated according to step 2 and each pixel
Change index Zx,y, further structure figure cuts the side of model, and specific construction method is as follows：
It is assumed that the side number that is built in total after (1) the end of the step is E, by newly build while be connected to previously built while it
Afterwards, specifically structure rule is：It is assumed that current pixel point i0<=i<=H*W-1Coordinate for (x, y), quantization index Zx,y, then add
Side be<i,E+Zx,y>；
(3) according to step 1 demarcate drawing of seeds picture, by the prospect of mark, background dot only with source point, meeting point S, T-phase connect
It connects, structure figure cuts the S-link and T-link of model；
(4) model is cut using maximum-flow algorithm solution figure, obtains a bianry image, the prospect being respectively partitioned into and the back of the body
Scape.
Step 5: merge not connected region, so as to obtain the prospect of image, background and zone of ignorance ternary diagram.
Two pieces of maximum regions are taken out, if there is other small cut zone inside every piece of region, then merges and does not connect
Same region.If zonule appears in background area, zonule is merged into background area, otherwise zonule appears in prospect
Zonule is then merged into foreground area by region.Merge connected component and belong to basic digital image processing techniques, do not go to live in the household of one's in-laws on getting married here
It states.Foreground area border is found out, opposite small range is belonged into the pixel of prospect and relatively large point range category along border
In the pixel of background ternary diagram is formed labeled as zone of ignorance.Given image is divided into prospect, the back of the body by ternary diagram Trimap
Scape and zone of ignorance to be asked.Ternary diagram is marked as prospect fI=0...F, background bI=0...B, zone of ignorance (region i.e. to be scratched)
uI=0...U, wherein F+B+U=H*W, H are picture altitude, and W is picture traverse.
Step 6: optional step, is extended ternary diagram according to the distribution of color of image, Dai Kou areas are further reduced
Domain：
According to certain Rule Extended image zone of ignorance uI=0...U, specific method is：
To each point of zone of ignorance in ternary diagram ui, calculate centered on its position (x, y), radius is r (5<=
r<=15, neighborhood Ω 10) is taken hereiniInterior pixel I (x', y') and point uiBetween present position (x, y) pixel I (x, y) color
Distance, if less than given threshold value (such as take the threshold value be 2), and the pixel belongs to prospect f (or background b).Then should
Pixel u expands to prospect f (or background b).By this extended operation, ternary diagram is further changed to background fI=0...F', background
bI=0...B', zone of ignorance uI=0...U'。
Color distance calculation formula is:
Step 7: calculating image gradient, image is sampled according to image gradient, to zone of ignorance, each point obtains
One group of prospect, background candidate point set：
1st, to each pixel II=0...H*W(R, G, B) extracts luminance channel according to the following formula：
L=0.299*R+0.587*G+0.114*B (10)
2nd, the gradient of image is calculated according to luminance channel：
Gx(x, y)=L (x+1, y)-L (x, y) (11)
Gy(x, y)=L (x, y+1)-L (x, y) (12)
3rd, to each pixel u in zone of ignoranceiIt is sampled according to its gradient direction and tangential direction, altogether four direction,
Each direction at least respectively samples a background dot of foreground point one, can so sample most 16 points as prospect, background sample
This point set
Four sample directions are respectively：
θ1=atan2 (Gy(x,y),Gx(x,y)) (13)
θ2,θ3,θ4θ can be passed through1To determine.It can be sampled if sample direction determines according to following rule,
Middle sampling step length is：
StepSize=min (1/sin (θi),1/cos(θi)) (14)
Coordinate update method is：
Y'=y+stepSize*sin (θi) (15)
X'=x+stepSize*cos (θi) (16)
Step 8: a pair of optimal prospect, background are selected each point in zone of ignorance from candidate point set
Point calculates a four-tuple.
To each point of zone of ignorance ui, the prospect that is sampled from previous step, background sample point setMiddle selection
Optimal a pair of of prospect, background dot fi、bi, specific method is as follows：
Selection is so that following metric relation gi,jMinimum a pair
Various computational methods are as follows in formula (17)：
M represents the color distortion degree of the prospect sampled, background dot and current pixel in formula, and N is then represented in its certain model
Enclose degreeof tortuosity in total in neighborhood.Good sampled point, it should possess minimum color distortion.
Color distortion is calculated in front to sampling the influence of point mass, but only considers that colouring information is inadequate, is gone back
The spatial positional information of pixel must be combined.E represents the line by image gradient along current point to sampled point in formula
Integration, if can obtain larger value across edge on path of integration, otherwise obtains smaller value.The background usually sampled
Point is more than foreground point, PFiFor calculating the probability that current pixel belongs to foreground point.It is to pass throughValue correction
PFiLater measurement current pixel point belongs to the probability of foreground point.
For unknown point and foreground point, the color distance of background dot, can be calculated by (9) formula,
It is that must be fulfilled for should very little with the space length of current point because of sampled point preferably that both this, which is added in formula (17),.Except from time
It selects and suitable f is selected in sampling point seti, bi, it is also necessary to further calculate the local variance sigma of unknown pointf 2,σb 2:
In this way, to each unknown point, a four-tuple can be obtained
Step 9: optionally, further optimization samples obtained prospect, background dot, calculates a new four-tuple：
It is suitable to be selected from candidate's sampled pointAfterwards, it is also necessary to further to selectingOptimization, tool
Body optimization process is as described below：
To each unknown point, centered on the point, k (1 in its 5 × 5 neighborhood unknown point is selected<=k<3)=5 take herein
It is a that there is minimum Mi(fj,bj) value four-tupleIts average value is obtainedAnd according to
The following formula calculates new four-tuple：Wherein the value of the confidenceRepresent what is calculatedBelong to the conviction of current pixel.Bigger, expression calculates more accurate, and finally stingy figure effect is got over
It is good.
Step 10: optionally, the new four-tuple that step 9 calculates is smoothed, so as to finally obtain α figures.
Calculate new four-tuple (fi,bi,αi,confidencei) after, it is also necessary to four-tuple is carried out smoothly, smoothly
Calculating process it is as follows：
G is Gaussian function in formula (31), remaining calculating is simple weighted average.
Here to calculating the value of the confidence of each pixelCarry out smooth filtering.
In formula：TuRepresent graph region to be scratched,
So as to finally obtain alpha figures.
In general, after the calculating of this step is completed, it is possible to obtain scratching figure effect well.It is but special for texture
When complicated, can seem very jagged sense, scheme for the α acquired, can be expressed as an optimization problem, but meeting again
Larger time overhead is brought, therefore the present invention increases a step, uses the optimal α values of fast algorithm approximate calculation：
Selection goes the approximate solution optimization problem using a kind of mode of filtering, uses Steerable filter (guidefilter)
Image is filtered, obtains the approximate solution of optimal α figures.
Step 11: by obtained optimal α scheme with the background colour (red, blueness, white etc.) specified and Original Photo piece according to
The mode of alpha mixing is mixed, you can obtains corresponding certificate photo image.
A preferred embodiment of the present invention has shown and described in above description, but as previously described, it should be understood that the present invention
Be not limited to form disclosed herein, be not to be taken as the exclusion to other embodiment, and available for various other combinations,
Modification and environment, and above-mentioned introduction or the technology or knowledge of association area can be passed through in the scope of the invention is set forth herein
It is modified.And changes and modifications made by those skilled in the art do not depart from the spirit and scope of the present invention, then it all should be in this hair
In the protection domain of bright appended claims.
Claims (7)
1. a kind of certificate photo generation method split based on image and scratch figure, specifically includes following step：The first step, mark
Go out the foreground and background of image, obtain drawing of seeds picture；The portrait that wherein prospect is face, hair and clothes for upper half of body form,
Background is not cover other regions of portrait；Three passages of drawing of seeds picture (R, G, B) are quantized into histogram by second step, are obtained
Go out the number for the Bin that drawing of seeds picture is actually used；3rd step, structure Image Segmentation Model to the histogram that second step obtains into
Row segmentation, foreground picture and Background after being split；4th step mixes obtained foreground picture with the background colour specified
It closes, obtains corresponding certificate photo image；
Wherein the 3rd step further comprises：
Color of image and space length are calculated, the side of model is cut according to color of image and space length structure figure；
The number of Bin of actual use and quantization index Zx, the y of each pixel calculated according to second step, further
Structure figure cuts the side of model；
According to the first step demarcate drawing of seeds picture, by the prospect of mark, background dot only with source point, meeting point S, T-phase connection, structure figure
Cut the S-link and T-link of model；
Model is cut using maximum-flow algorithm solution figure, obtains a bianry image, the foreground and background being respectively partitioned into；
It is wherein further comprising the steps of before the 4th step after the 3rd step：
Merge not connected region, obtain the prospect, background and zone of ignorance ternary diagram of image；
Image gradient is calculated, image is sampled according to image gradient, each point obtains one group of prospect, the back of the body to zone of ignorance
Scape candidate point set；
The optimal prospect of a pile, background dot from candidate point set are selected to each point in zone of ignorance, calculate one
A four-tuple.
2. split and scratch the certificate photo generation method of figure based on image as described in claim 1, it is characterised in that the method
It further includes and ternary diagram is extended according to the distribution of color of image, further reduce zone of ignorance.
3. split and scratch the certificate photo generation method of figure based on image as described in claim 1 or 2, it is characterised in that described
Method further includes prospect, the background dot that optimization sampling obtains, and calculates a new four-tuple.
4. split and scratch the certificate photo generation method of figure based on image as claimed in claim 3, it is characterised in that the method
It further includes and is smoothed to calculating four-tuple.
5. split and scratch the certificate photo generation method of figure based on image as claimed in claim 4, it is characterised in that the method
It further includes using the optimal foreground image of fast algorithm approximate calculation.
6. split and scratch the certificate photo generation method of figure based on image as claimed in claim 5, it is characterised in that described image
Image is split using maximum-flow algorithm in segmentation.
7. split and scratch the certificate photo generation method of figure based on image as claimed in claim 6, it is characterised in that described to four
The mode that tuple is smoothed is to be smoothed using Gaussian function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510022078.8A CN104504745B (en) | 2015-01-16 | 2015-01-16 | A kind of certificate photo generation method split based on image and scratch figure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510022078.8A CN104504745B (en) | 2015-01-16 | 2015-01-16 | A kind of certificate photo generation method split based on image and scratch figure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104504745A CN104504745A (en) | 2015-04-08 |
| CN104504745B true CN104504745B (en) | 2018-05-25 |
Family
ID=52946139
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510022078.8A Active CN104504745B (en) | 2015-01-16 | 2015-01-16 | A kind of certificate photo generation method split based on image and scratch figure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104504745B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881287B (en) * | 2015-05-29 | 2018-03-16 | 广东欧珀移动通信有限公司 | Screenshot method and device |
| CN105184787B (en) * | 2015-08-31 | 2018-04-06 | 广州市幸福网络技术有限公司 | A kind of license camera and method for portrait scratch figure automatically |
| CN105120167B (en) * | 2015-08-31 | 2018-11-06 | 广州市幸福网络技术有限公司 | A kind of license camera and license image pickup method |
| CN106952270A (en) * | 2017-03-01 | 2017-07-14 | 湖南大学 | A kind of quickly stingy drawing method of uniform background image |
| CN107123088B (en) * | 2017-04-21 | 2019-09-13 | 山东大学 | A kind of method of automatic replacement photo background color |
| CN108156370A (en) * | 2017-12-07 | 2018-06-12 | Tcl移动通信科技（宁波）有限公司 | By the use of local picture as the photographic method of background, storage medium and mobile terminal |
| CN108053366A (en) * | 2018-01-02 | 2018-05-18 | 联想(北京)有限公司 | A kind of image processing method and electronic equipment |
| CN108965707B (en) * | 2018-07-23 | 2020-11-13 | 金展科技（佛山）有限公司 | Automatic bottom-removing shooting system |
| CN109389611A (en) * | 2018-08-29 | 2019-02-26 | 稿定（厦门）科技有限公司 | The stingy drawing method of interactive mode, medium and computer equipment |
| CN109191486A (en) * | 2018-09-12 | 2019-01-11 | 广州粤创富科技有限公司 | A kind of pet image partition method and electronic equipment |
| CN111476146B (en) * | 2020-04-03 | 2023-04-07 | 北京金沙江科技有限公司 | Mobile-terminal-based academic record certification standardized processing method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2658727C (en) * | 2006-07-25 | 2017-10-24 | Humaneyes Technologies Ltd. | Interactive segmentation of images with single scribbles |
| US8175384B1 (en) * | 2008-03-17 | 2012-05-08 | Adobe Systems Incorporated | Method and apparatus for discriminative alpha matting |
| KR101670282B1 (en) * | 2009-02-10 | 2016-10-28 | 톰슨 라이센싱 | Video matting based on foreground-background constraint propagation |
| CN102917175A (en) * | 2012-09-13 | 2013-02-06 | 西北工业大学 | Sheltering multi-target automatic image matting method based on camera array synthetic aperture imaging |
-
2015
- 2015-01-16 CN CN201510022078.8A patent/CN104504745B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| Interactive graph cuts for optimal boundary & region segmentation of objects in N-D images Computer Vision;Boykov Y Y等;《ICCV 2001. Proceedings. Eighth IEEE International Conference on. IEEE》;20010714;第107-109页 * |
| 自动静态图像前景提取;赵瑞霖;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130915(第9期);摘要，第1－2页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104504745A (en) | 2015-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104504745B (en) | A kind of certificate photo generation method split based on image and scratch figure | |
| CN107862698B (en) | Light field foreground segmentation method and device based on K mean cluster | |
| CN107516319B (en) | High-precision simple interactive matting method, storage device and terminal | |
| US9483835B2 (en) | Depth value restoration method and system | |
| JP4898800B2 (en) | Image segmentation | |
| WO2017054314A1 (en) | Building height calculation method and apparatus, and storage medium | |
| Dai et al. | Single underwater image restoration by decomposing curves of attenuating color | |
| Gao et al. | Sand-dust image restoration based on reversing the blue channel prior | |
| CN104811684B (en) | A kind of three-dimensional U.S. face method and device of image | |
| KR20130045242A (en) | Method and device for generating morphing animation | |
| CN111563908B (en) | Image processing method and related device | |
| CN108564528A (en) | A kind of portrait photo automatic background weakening method based on conspicuousness detection | |
| CN106373128B (en) | Method and system for accurately positioning lips | |
| CN109064418A (en) | A kind of Images Corrupted by Non-uniform Noise denoising method based on non-local mean | |
| Vu et al. | Efficient hybrid tree-based stereo matching with applications to postcapture image refocusing | |
| CN108320294B (en) | Intelligent full-automatic portrait background replacement method for second-generation identity card photos | |
| CN108810415B (en) | Focusing method based on quantum particle swarm optimization algorithm | |
| CN109525786A (en) | Method for processing video frequency, device, terminal device and storage medium | |
| CN111242074B (en) | Certificate photo background replacement method based on image processing | |
| US9786055B1 (en) | Method and apparatus for real-time matting using local color estimation and propagation | |
| CN105740867B (en) | The selection method of image texture window shape and scale | |
| AU2016273979A1 (en) | System and method for adjusting perceived depth of an image | |
| CN114677393B (en) | Depth image processing method, depth image processing device, image pickup apparatus, conference system, and medium | |
| CN108961258B (en) | Foreground image obtaining method and device | |
| CN106327500B (en) | Depth information acquisition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP02 | Change in the address of a patent holder |
Address after: 610015 floor 13, building 1, No.1268, middle section of Tianfu Avenue, Chengdu high tech Zone, China (Sichuan) pilot Free Trade Zone, Chengdu Patentee after: Chengdu PinGuo Digital Entertainment Ltd. Address before: 610041 C12-16 building, Tianfu Software Park, hi tech Zone, Sichuan, Chengdu Patentee before: Chengdu PinGuo Digital Entertainment Ltd. |
|
| CP02 | Change in the address of a patent holder |