CN108537126A - A kind of face image processing system and method - Google Patents
A kind of face image processing system and method Download PDFInfo
- Publication number
- CN108537126A CN108537126A CN201810205659.9A CN201810205659A CN108537126A CN 108537126 A CN108537126 A CN 108537126A CN 201810205659 A CN201810205659 A CN 201810205659A CN 108537126 A CN108537126 A CN 108537126A
- Authority
- CN
- China
- Prior art keywords
- image
- face
- image processing
- customer
- hair style
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0641—Shopping interfaces
- G06Q30/0643—Graphical representation of items or shoppers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/40—Filling a planar surface by adding surface attributes, e.g. colour or texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/60—Editing figures and text; Combining figures or text
Abstract
The invention discloses a kind of face image processing system and method, the system comprises：Model's image storage module, for storing model's image；Man face image acquiring module, for acquiring facial image；Image transmission module is used for transmission facial image to image processing module；Image processing module, for carrying out face image synthesis；Image display includes the facial image of acquisition, model's image of recommendation and final composograph for showing.The method mainly detection input picture and the human face region in reference picture simultaneously extract human face characteristic point, and according to identical rule respectively to two facial images progress triangulations.System and method through the invention can be such that image capture module is detached with image processing module realization, convenient for selecting best acquisition position, while enabling a customer to see collected facial image in real time, select suitable human face expression, best picture-taking position.
Description
Technical field
The present invention relates to a kind of image processing system and method, especially a kind of system for face image processing and
Method.
Background technology
Image processing techniques is applied to include many fields such as medical treatment, military, production.Image processing techniques is in face figure
As the application in identification, people is allow more easily to obtain relevant information in many fields, to be made more to correlation circumstance
Accurately judge.
Under many scenes, in hairdressing place, customer wishes before haircut that hair style, there are one straight if appropriate for oneself
The judgement of sight, so that the selection for being more suitable for oneself can be made.In addition, this method by face image synthesis is protected in privacy
Shield, virtual fitting, amusement and leisure etc. also all have broad application prospects.
Major defect existing for currently available technology is：
First, composograph authenticity can be caused not high if brightness of image to be synthesized, contrast, tone are inconsistent,
To which useful information can not be obtained.
Second is that when applying in above-mentioned barber shop's scene, since hair style can cause to block to human face's face, customer's
Facial characteristics can not be seamless applying with the hair style of model, causes image effect unnatural so that Customer Experience is deteriorated.
Third, the prior art can not change the face mask of composograph, popular says, provided that model be round
Face, but the shape of face side of the being face of customer, existing synthetic technology can only make composograph be still round face.
In the application of the paper and realization delivered, what is had can change the human face five-sense-organ of reference picture, final to present
Effect be exactly the face of composograph be the face feature of customer, but three above-mentioned problems are still existing.
" the Face Swapping under Large Pose Variations of Lin Y, Wang S, Lin Q et al.: A
3D Model Based Approach " are (in IEEE, International Conference on Multimedia and
Expo, 2012) it is middle using the method for establishing 3D head models (3D morphable models, 3dMM), due to technology
It is immature, the face feature of part is can be only generated, this can cause last synthetic effect untrue, and be based on this 3D moulds
The method of type can all devote a tremendous amount of time, and cannot meet most application scenarios.
Invention content
The technical problem to be solved in the present invention is to provide a kind of face image processing system and methods, and facial image is closed
At effect have higher authenticity, for customer provide more intuitively, more accurately simulate effect.
The technical proposal of the invention is realized in this way：
A kind of face image processing system, including：Model's image storage module, for storing model's image；Facial image
Acquisition module, for acquiring facial image；Image transmission module is used for transmission facial image to image processing module；At image
Module is managed, for carrying out face image synthesis；Image display includes the facial image acquired, the mould of recommendation for showing
Special image and final composograph.
Preferably, the system also includes data analysis unit, for analyzing customer data information, the number of Main Analysis
According to the hair style for including customer's age, gender and suitable customer, it is suitble to the hair style of oneself to carry out Primary Stage Data standard for customer's selection
It is standby.
Preferably, the system also includes visible user interface, the visible user interface include for taking pictures and
The button of image selection allows user to take pictures to customer in proper moment and select photo.
It is pushed away preferably, the visible user interface further includes the data progress hair style analyzed according to customer's appearance
The option button recommended.
Preferably, the visible user interface further includes the hair style model for checking databases storage for customer
Option button wishes to attempt other hair styles, option permission user's hand in database after user sees the hair style of recommendation
The hair style model that dynamic selection is liked oneself, and generate the image of oneself hair style identical as model using synthesis button.
Preferably, described image transmission module includes the network element for wireless data transmission, which allows user
The wirelessly remote control camera in same LAN.
Preferably, described image processing module includes image pre-processing unit, for being carried out to collected custom image
Pretreatment, the pretreatment includes image gray processing, histogram equalization and filtering operation.
Preferably, described image processing module further includes feature extraction unit, it to be used for customer face region detection and people
Face characteristic point extracts, and the human face characteristic point is a series of point that can reflect face feature of predefined, main point
It is distributed in human face five-sense-organ profile.
Preferably, described image processing module further includes the image composing unit for composograph, the unit for pair
Image between the custom image and hair style model of acquisition synthesizes, make composograph face and shape of face all with customer
Unanimously, and hair style is consistent with the hair style of model.
A kind of face image processing process, the method be suitable for above-mentioned technical proposal described in any system, including with
Lower step：
S1 detects the human face region in input picture and reference picture, and extracts human face characteristic point respectively；
S2 carries out triangle to two facial images respectively based on the human face characteristic point extracted in S1 according to identical rule
Subdivision；
S3, the affine change between the correspondence triangle that calculating input image to reference picture is obtained through the triangulation
It changes, and color filling is carried out to the triangle in reference picture, obtain an intermediate image；
S4, from the extraction face area-of-interest of intermediate image described in S3 (ROI)；
S5 makes face area-of-interest (ROI) mask images, and the mask images are for handling due to input picture
With reference picture heterochromia caused by the unnatural problem of composograph；
S6 completes the color correction of composograph by face mask images so that the color perfection of composograph
Degree improves authenticity.
The beneficial effects of the present invention are：
1. image capture module can be made to realize with image processing module using the system architecture of invention to detach, convenient for selection
Best acquisition position, while enabling a customer to see collected facial image in real time, suitable human face expression is selected, most
Good picture-taking position.
2. considerably increasing the degree of freedom in space using the transmission mode of long distance wireless, and image capture device volume
It is small and exquisite, it can be used as handheld device, customer can be with hand-held image collecting device, as long as collecting device and image procossing mould
Block is in same LAN, and the system can be achieved with the picture real-time Transmission between collecting device and processing module, can make bat
Picture is observed simultaneously according to personnel and customer, and adjusts photo angle in real time.
3. customer can be made to manually select the hair oneself liked in database using the visualization interface with function choosing-item
Pattern is special, and generates the image after oneself hair style identical as model using synthesis button, considerably increases grasping for user
The property made provides the selection of more diversification for customer.
Description of the drawings
Attached drawing 1 is the structural schematic diagram of system of the present invention；
Attached drawing 2 is system user terminal application scenario diagram of the present invention；
Attached drawing 3 is that Face datection and human face characteristic point extract schematic diagram in the embodiment of the present invention；
Attached drawing 4 is the first schematic diagram to image triangulation in the embodiment of the present invention；
Attached drawing 5 is the second schematic diagram to image triangulation in the embodiment of the present invention；
Attached drawing 6 is the face mask images made in the embodiment of the present invention；
Attached drawing 7 is image composition algorithm flow chart of the present invention；
Attached drawing 8 is image composition algorithm embodiment flow chart of the present invention.
Specific implementation mode
The present invention is described in further details with reference to the accompanying drawings and examples：
As shown in Fig. 1, a kind of face image processing system, including：
Model's image storage module, for storing model's image；
Man face image acquiring module, for acquiring facial image；
Image transmission module is used for transmission facial image to image processing module；
Image processing module, for carrying out face image synthesis；
Image display includes the facial image of acquisition, model's image of recommendation and final synthesis for showing
Image.
Further, the system also includes data analysis units, for analyzing customer data information, Main Analysis
Data include the hair style at customer's age, gender and suitable customer, are suitble to the hair style of oneself to carry out Primary Stage Data for customer's selection
Prepare.
Further, the system also includes visible user interface, the visible user interface includes for taking pictures
With the button of image selection, user is allowed to take pictures to customer in proper moment and select photo, and according to Gu
The data that objective appearance is analyzed carry out the option button of hair style recommendation and check the hair style mould of databases storage for customer
Special option button wishes that other hair styles of trial, the option allow user in database after user sees the hair style of recommendation
The hair style model oneself liked is inside manually selected, and the image of oneself hair style identical as model is generated using synthesis button.
Further, described image transmission module includes the network element for wireless data transmission, which allows to use
Family wirelessly remote control camera in same LAN.
Further, described image processing module includes image pre-processing unit, for collected custom image into
Row pretreatment, the pretreatment includes image gray processing, histogram equalization and filtering operation.
Further, described image processing module further includes feature extraction unit, for customer face region detection and
Human face characteristic point extracts, and the human face characteristic point is a series of point that can reflect face feature of predefined, mainly
It is distributed in human face five-sense-organ profile.Further, described image processing module further includes the image synthesis list for composograph
Member, the unit are synthesized for image between the custom image and hair style model of acquisition, make the face of composograph with
And shape of face is all consistent with customer, and hair style is consistent with the hair style of model.
Further, described image processing module further includes the image composing unit for composograph.The unit can
It is synthesized with the image between hair style model for the custom image of acquisition, the effect finally presented is exactly the five of composograph
Official and shape of face are all consistent with customer, but hair style is consistent with the hair style of model.This is the core missions of system.
In the present embodiment, as shown in attached drawing 1,2, man face image acquiring module is used for Image Acquisition of customer when into shop
And the acquisition that hair style selection is more accurate facial image is carried out, the image of image transmission module acquisition uses wireless side
Formula is transmitted between image capture module and image processing module, image processing module completion include the data analysis of customer's appearance,
Multiple system core functions such as model's hair style is recommended, image data synthesizes, user people of the image display arrangement for acquisition
Face image is shown, progress is shown, composograph is shown, user interface is shown for image synthesis.Specifically, described image acquires
Between module and image processing module by the way of wireless remote transmission, image processing module accounts for fixed dimension in space,
And can keep stable in the absence of external forces, suitable for being positioned over fixed space position；Image processing module is shown with image
Between module by the way of wired connection, therefore image processing module and image display are placed on same place space bit
It sets.Since image processing module volume is too big, it is not suitable as mobile device use, therefore image capture module can be with figure
As processing module separation use, convenient for selecting best acquisition position, while enabling a customer to see collected face in real time
Image selects suitable human face expression, best picture-taking position.Simultaneously because customer, which cannot constantly take into account, takes pictures and keeps dynamic
Make, therefore specific action of taking pictures is arranged in image processing module, is completed by another people, also can in image processing module
The concrete condition of real-time display acquisition.The transmission mode of long distance wireless considerably increases the freedom in space, and Image Acquisition
Equipment volume is small and exquisite, can be used as handheld device, and customer can be with hand-held image collecting device, as long as collecting device and figure
Picture processing module is in same LAN, and the system can be achieved with the picture between collecting device and processing module and pass in real time
It is defeated, allow the personnel of taking pictures to observe picture simultaneously with customer, and adjust photo angle in real time.When the image of customer is collected mould
Block captures, and processing module is transferred to by way of wireless transmission, image processing module first can analyze the personal information of customer
The precise information based on customer's essential information is obtained in order to recommend the hair style for being more suitable for customer, at this time image display
It can show the hair style of the suitable customer of recommendation, the Gu that image synthesis unit can be directed to model's image of hair style and transmit
Objective image carries out image synthesis and carries out the displaying of composograph in display module.On the one hand this function is conducive to save and care for
The objective time need not be manually entered information because this part of functions is entirely by systematic automatic operation；On the other hand it cares for
Visitor can improve the consumption experience of customer with the appropriateness of direct feel hair style and oneself.
As shown in Fig. 2, handheld device (recommendation uses smart mobile phone in the present invention) front camera can be used in customer
Carry out information collection, shooting picture can on the screen of handheld device real-time display, while user is in the aobvious of image display
Also the picture captured by customer can be received in display screen by way of wireless transmission, action of taking pictures is by user's control, when thinking
Suitable picture is taken, the Esc keys that can be pressed on lower keyboard are shot.Collected custom image passes through wireless transmission
Mode is received by image processing module, analyzes the surface information of customer first, matches model's figure all in memory module
Picture has selected the hair style of 4 most suitable customers in embodiment, on a display screen the effect after displaying synthesis, and customer is also
Hair style model all in library can be manually selected to carry out image synthesis.
Image processing module described in the present embodiment includes being extracted for customer face region detection and human face characteristic point
Feature extraction unit.As shown in attached drawing 3,4,5, what human face characteristic point referred to predefined can reflect face feature
A series of point, be distributed mainly on human face five-sense-organ profile.Feature extraction unit can provide the tool where customer face region
The specific distributing position of body position and human face characteristic point.The distributing position of customer's human face region position and characteristic point will be people
Face information data is analyzed and the important preparation data of face image synthesis.
As shown in Fig. 7, a kind of face image processing process, the method are suitable for any described in above-described embodiment
System includes the following steps：S1 detects the human face region in input picture and reference picture, and extracts face characteristic respectively
Point；S2 carries out triangulation to two facial images respectively based on the human face characteristic point extracted in S1 according to identical rule；
S3, the affine transformation between the correspondence triangle that calculating input image to reference picture is obtained through the triangulation, and to ginseng
The triangle examined in image carries out color filling, obtains an intermediate image；S4 extracts face from intermediate image described in S3
Area-of-interest (ROI)；S5, make face area-of-interest (ROI) mask images, the mask images for handle due to
The unnatural problem of composograph caused by input picture and reference picture heterochromia；S6 is completed by face mask images
The color correction of composograph so that the color of composograph is perfect excessively, improves authenticity.
More specific embodiment of the method, as shown in Fig. 8：
Step 101 is the acquisition of customer's facial image, and the image collected is called input picture.Input picture is most
Whole composograph provides face and face mask information.
Step 102 is the pretreatment operation of the input picture, includes mainly various be filtered, it is intended to improve image
Quality.
Step 103 is Face datection, and Face datection algorithm uses the HAAR that existing computer vision library OPENCV is provided
(Haar-like features) algorithm is detected face, its return value is a square region for including human face region, tool
The mathematical notation of body is this box top left corner apex coordinate and rectangular length and width.
Step 104 is facial feature points detection, needs to detect the key point of face to determine the position of face feature.It adopts
With Kazemi, Vahid, Josephine Sullivan et al. are in paper " One Millisecond Face Alignment
With An Ensemble Of Regression Trees " (In IEEE, Conference on Computer Vision
And Pattern Recognition (CVPR), 2014) in the improved ASM methods that propose, as a result can obtain 68 people
Face characteristic point, they sketch the contours of eyebrow, eyes, nose, face and the face mask of face, and attached drawing 3 illustrates the people of image
Face detects and characteristic point testing result.
Input picture is aligned by step 105 with the face in reference picture, including scaling, rotation process, in order to make two
Facial size, the angle of image are consistent, and select the vector between the 40th and the 43rd characteristic point for basic rotating vector,
The angle between the vector of two images is sought, and rotation process is done to input picture so that two image facial angles are kept
Unanimously, it then selects the vector between the 1st and the 17th characteristic point as basic scale vectors, scaling behaviour is done to input picture
Make.After face alignment, the facial size and angle of input picture and reference picture can be consistent, input picture rotation
After turning, human face characteristic point will also change therewith.This is highly important to work later.
Step 106 is to carry out triangulation to input picture.
It needs to do triangulation to input picture and reference picture respectively, first has to increase several characteristic points, for input
Image is increased by 3 characteristic points, is mathematically indicated using following formula：
For reference picture, increase by 7 characteristic points：
Also 4 characteristic points are the coordinates at four angles of reference picture.
WhereinIndicate the coordinate of the 69th characteristic point of input picture, it includes two values, indicates thinking for characteristic point
X, y-coordinate value.Indicate the coordinate of the ith feature point of reference picture, α is a coefficient, oneself can be chosen, herein
Enable α=1.2.Intuitively these three characteristic points are located at the top in the centre position and two eyebrows of face eyes, separately
Outside in order to make the shape of the face after changing face be consistent with input picture, the human face characteristic point of reference picture is changed, from number
It is so indicated on：
Triangulation as shown in Figure 5 is carried out to reference picture first, this operation will carry out twice, be directed to spy respectively
Reference picture before sign point changes and after changing.It is a series of right two-by-two to have been obtained on two images after triangulation
The triangle surface answered.
Calculate affine transformation so that each triangular apex of image is mapped to characteristic point change before characteristic point changes
Image later corresponds in triangular apex.
Affine transformation includes：
A) (linear transformation) is rotated
B) translation (vector adds)
Zoom operations (linear transformation)
Three points are assured that an affine transformation in image, and affine change is indicated usually using 2 × 3 matrixes
It changes：
Matrix A, B can be to bivectorsIt converts, so following form can also be expressed as：
Or
T=M [x, y, 1]T (10)
T is exactly vectorial X by the vector after affine transformation M.In step 112, it is known that vectorial X and vector T,
That to be solved is transformation matrix M.
According to3 pairs of points are chosen in two images respectively, so that it may to obtain six equations, solve
All values in matrix M.
Then with the affine transformation calculated by all pixels in the unchanged reference picture intermediate cam shape of characteristic point
In reference picture after point deformation to characteristic point change.By this step can obtain face do not change but face
It is called second level reference picture by the shape image consistent with input picture.The characteristic point of second level reference picture is
The characteristic point of the reference picture of change.
Second step carries out triangulation to input picture, and the rule of triangulation is as shown in figure 4, because only need to input
The face feature of image, so triangulation need not all be carried out to whole input picture.Input is can be seen that from Fig. 4 and Fig. 5
Face's triangulation of image and second level reference picture is consistent, therefore the triangle that they are obtained is also to correspond
's.Likewise, repeating the operation of previous step, affine transformation is calculated, but is specifically by three of each triangle in input picture
A vertex is mapped in the reference picture of the second level.An image, the facial contours of image and five can be obtained after the completion of this step
Official's feature is all consistent with input picture, and other parts are consistent with reference picture, this image is become third level reference picture,
Its characteristic point is consistent with the characteristic point of second level reference picture.
Step 107~111 are to be directed to reference picture, and specific embodiment is consistent with step 101~105,
Step 112 hereinbefore illustrates.
An intermediate image, the corresponding third level reference picture can be obtained after the completion of step 112.In order to certainly
Dynamic realization institute is functional, and the ROI of face is extracted in step 113 or based on characteristic point is detected.In paper, choosing
Take 7 feature point groups at convex closure as ROI.Shown in the following formula of seven points：
WhereinIndicate first feature point coordinates of ROI, β is a parameter that can be manually selected, and is enabled in paper
β=0.05.ROI is extracted from third level reference picture, includes most of feature of face in ROI.Colour correction is for calculating
Particularly significant for method, the authenticity and the color compactness relationship of ROI and reference picture for result of changing face are very close, then
He is fitted in the reference picture of the second level, in order to realize the color bumpless transfer of ROI and second level reference picture, is used
The function provided in the libraries OPENCV realizes the slitless connection of human face's color, and image is made to seem more true to nature, natural.
In addition to that face ROI can be calculated by characteristic point, the present invention also provides a kind of making face mask images
Method.
As shown in Fig. 6, mask images are also a kind of method for extracting face ROI, it is a bianry image,
Only two kinds of colors of black and white, white portion correspond to face ROI, and black portions are that information abandons part.For the hair stored in library
Pattern is special, it is only necessary to make a mask images, can repeat to have invoked in program afterwards.
The method for making mask images is easily understood very much, and the method provided is to select mask images by left mouse button
The borderline point of white portion so that the point of selection can include the face information of model, but not destroy hair style information.
The number of the point of selection and position can freely control.
Step 114 is that facial characteristics exchanges, this step is exactly that the ROI of extraction is fitted to the second level reference picture
In.When handling this problem there are one situation it is to be appreciated that being exactly the position of ROI fittings.Because everyone face is several
What is different, if the position selection inaccuracy of ROI can cause face to be subjected to displacement when being changed face, manually
It is to change face to operate each time all to select suitable position that can waste the plenty of time, solves the problems, such as this in the present invention.It is first
The minimum rectangle of an encirclement ROI is first established, and seeks the coordinate of this rectangular central point, uses C1Then it is indicated,
In the reference picture of the second level, based on ROI feature point is sought, a minimum rectangle for surrounding them is established, rectangle is sought
Center position coordinate C2, make C when finding the positions ROI1With C2It overlaps, preferable effect can be obtained.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited to
This, any system and method described through the invention carries out face image processing, to realize that customer's hair style simulates matched system
And method mentality of designing belongs to the protection domain of the technology of the present invention design, any one skilled in the art exists
The invention discloses technical scope in, according to the technique and scheme of the present invention and its design be subject to equivalent substitution or change, should all contain
Lid is within protection scope of the present invention.
Claims (10)
1. a kind of face image processing system, which is characterized in that including：
Model's image storage module, for storing model's image；
Man face image acquiring module, for acquiring facial image；
Image transmission module is used for transmission facial image to image processing module；
Image processing module, for carrying out face image synthesis；
Image display includes the facial image of acquisition, model's image of recommendation and final composograph for showing.
2. image processing system according to claim 1, it is characterised in that：The system also includes data analysis unit,
For analyzing customer data information, the data of Main Analysis include the hair style at customer's age, gender and suitable customer, are customer
Selection is suitble to the hair style of oneself to carry out Primary Stage Data preparation.
3. image processing system according to claim 1, it is characterised in that：The system also includes visual users circle
Face, the visible user interface include the button for taking pictures with image selection, allow user proper moment to customer into
Row is taken pictures and is selected photo.
4. image processing system according to claim 3, it is characterised in that：The visible user interface further includes basis
The data that customer's appearance is analyzed carry out the option button of hair style recommendation.
5. image processing system according to claim 3, it is characterised in that：The visible user interface further includes being used for
Customer checks the option button of the hair style model of databases storage, wishes to attempt other after user sees the hair style of recommendation
Hair style, which allows user to manually select the hair style model oneself liked in database, and is generated using synthesis button
The image of oneself hair style identical as model.
6. image processing system according to claim 1, it is characterised in that：Described image transmission module includes for wireless
The network element of data transmission, the unit allow user's wirelessly remote control camera in same LAN.
7. image processing system according to claim 1, it is characterised in that：Described image processing module includes that image is located in advance
Unit is managed, for being pre-processed to collected custom image, the pretreatment includes image gray processing, histogram equalization
And filtering operation.
8. image processing system according to claim 1, it is characterised in that：Described image processing module further includes that feature carries
Take unit, extracted for customer face region detection and human face characteristic point, the human face characteristic point be predefined can
The a series of point for reflecting face feature, is distributed mainly on human face five-sense-organ profile.
9. image processing system according to claim 1, it is characterised in that：Described image processing module further includes for closing
At the image composing unit of image, image of the unit between the custom image and hair style model of acquisition synthesizes,
Keep face and the shape of face of composograph all consistent with customer, and hair style is consistent with the hair style of model.
10. a kind of face image processing process, which is characterized in that the method is suitable for claim 1-9 any claims
The system, includes the following steps：
S1 detects the human face region in input picture and reference picture, and extracts human face characteristic point respectively；
S2 carries out triangulation to two facial images respectively based on the human face characteristic point extracted in S1 according to identical rule；
S3, the affine transformation between the correspondence triangle that calculating input image to reference picture is obtained through the triangulation, and
Color filling is carried out to the triangle in reference picture, obtains an intermediate image；
S4, from the extraction face area-of-interest of intermediate image described in S3 (ROI)；
S5 makes face area-of-interest (ROI) mask images, and the mask images are for handling due to input picture and reference
The unnatural problem of composograph caused by image color difference；
S6 completes the color correction of composograph by face mask images so that the color of composograph is perfect excessively, improves
Authenticity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810205659.9A CN108537126B (en) | 2018-03-13 | 2018-03-13 | Face image processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810205659.9A CN108537126B (en) | 2018-03-13 | 2018-03-13 | Face image processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108537126A true CN108537126A (en) | 2018-09-14 |
| CN108537126B CN108537126B (en) | 2021-03-23 |
Family
ID=63484557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810205659.9A Active CN108537126B (en) | 2018-03-13 | 2018-03-13 | Face image processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108537126B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110503599A (en) * | 2019-08-16 | 2019-11-26 | 珠海天燕科技有限公司 | Image processing method and device |
| CN110543826A (en) * | 2019-08-06 | 2019-12-06 | 尚尚珍宝(北京)网络科技有限公司 | Image processing method and device for virtual wearing of wearable product |
| CN110555812A (en) * | 2019-07-24 | 2019-12-10 | 广州视源电子科技股份有限公司 | image adjusting method and device and computer equipment |
| CN110610456A (en) * | 2019-09-27 | 2019-12-24 | 上海依图网络科技有限公司 | Imaging system and video processing method |
| CN112102146A (en) * | 2019-06-18 | 2020-12-18 | 北京陌陌信息技术有限公司 | Face image processing method, device, equipment and computer storage medium |
| CN112769937A (en) * | 2021-01-12 | 2021-05-07 | 济源职业技术学院 | Medical treatment solid waste supervisory systems |
| US11074733B2 (en) | 2019-03-15 | 2021-07-27 | Neocortext, Inc. | Face-swapping apparatus and method |
| CN116228763A (en) * | 2023-05-08 | 2023-06-06 | 成都睿瞳科技有限责任公司 | Image processing method and system for eyeglass printing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404910A (en) * | 2006-03-23 | 2009-04-08 | 花王株式会社 | Hair style simulation image creating method |
| CN103646416A (en) * | 2013-12-18 | 2014-03-19 | 中国科学院计算技术研究所 | Three-dimensional cartoon face texture generation method and device |
| US8798321B2 (en) * | 2005-09-28 | 2014-08-05 | Facedouble, Inc. | Method and system for tagging an image of an individual in plurality of photos |
| CN105045968A (en) * | 2015-06-30 | 2015-11-11 | 青岛理工大学 | Hairstyle design method and system |
| CN105117445A (en) * | 2015-08-13 | 2015-12-02 | 北京建新宏业科技有限公司 | Automatic hairstyle matching method, device and system |
| CN105354411A (en) * | 2015-10-19 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Information processing method and apparatus |
| CN107741974A (en) * | 2017-10-09 | 2018-02-27 | 武汉轻工大学 | Aid in hairdressing method |
| CN107784134A (en) * | 2016-08-24 | 2018-03-09 | 南京乐朋电子科技有限公司 | A kind of virtual hair style simulation system |
-
2018
- 2018-03-13 CN CN201810205659.9A patent/CN108537126B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8798321B2 (en) * | 2005-09-28 | 2014-08-05 | Facedouble, Inc. | Method and system for tagging an image of an individual in plurality of photos |
| CN101404910A (en) * | 2006-03-23 | 2009-04-08 | 花王株式会社 | Hair style simulation image creating method |
| CN103646416A (en) * | 2013-12-18 | 2014-03-19 | 中国科学院计算技术研究所 | Three-dimensional cartoon face texture generation method and device |
| CN105045968A (en) * | 2015-06-30 | 2015-11-11 | 青岛理工大学 | Hairstyle design method and system |
| CN105117445A (en) * | 2015-08-13 | 2015-12-02 | 北京建新宏业科技有限公司 | Automatic hairstyle matching method, device and system |
| CN105354411A (en) * | 2015-10-19 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Information processing method and apparatus |
| CN107784134A (en) * | 2016-08-24 | 2018-03-09 | 南京乐朋电子科技有限公司 | A kind of virtual hair style simulation system |
| CN107741974A (en) * | 2017-10-09 | 2018-02-27 | 武汉轻工大学 | Aid in hairdressing method |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11074733B2 (en) | 2019-03-15 | 2021-07-27 | Neocortext, Inc. | Face-swapping apparatus and method |
| CN112102146A (en) * | 2019-06-18 | 2020-12-18 | 北京陌陌信息技术有限公司 | Face image processing method, device, equipment and computer storage medium |
| CN112102146B (en) * | 2019-06-18 | 2023-11-03 | 北京陌陌信息技术有限公司 | Face image processing method, device, equipment and computer storage medium |
| CN110555812A (en) * | 2019-07-24 | 2019-12-10 | 广州视源电子科技股份有限公司 | image adjusting method and device and computer equipment |
| WO2021012599A1 (en) * | 2019-07-24 | 2021-01-28 | 广州视源电子科技股份有限公司 | Image adjustment method, apparatus, and computer device |
| CN110543826A (en) * | 2019-08-06 | 2019-12-06 | 尚尚珍宝(北京)网络科技有限公司 | Image processing method and device for virtual wearing of wearable product |
| CN110503599A (en) * | 2019-08-16 | 2019-11-26 | 珠海天燕科技有限公司 | Image processing method and device |
| CN110503599B (en) * | 2019-08-16 | 2022-12-13 | 郑州阿帕斯科技有限公司 | Image processing method and device |
| CN110610456A (en) * | 2019-09-27 | 2019-12-24 | 上海依图网络科技有限公司 | Imaging system and video processing method |
| CN112769937A (en) * | 2021-01-12 | 2021-05-07 | 济源职业技术学院 | Medical treatment solid waste supervisory systems |
| CN116228763A (en) * | 2023-05-08 | 2023-06-06 | 成都睿瞳科技有限责任公司 | Image processing method and system for eyeglass printing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108537126B (en) | 2021-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108537126A (en) | A kind of face image processing system and method | |
| CN109690617B (en) | System and method for digital cosmetic mirror | |
| US9959453B2 (en) | Methods and systems for three-dimensional rendering of a virtual augmented replica of a product image merged with a model image of a human-body feature | |
| CN108305312B (en) | Method and device for generating 3D virtual image | |
| CN109840825A (en) | The recommender system of physical features based on user | |
| CN105354876B (en) | A kind of real-time volume fitting method based on mobile terminal | |
| JP4435809B2 (en) | Virtual makeup apparatus and method | |
| CN108765273A (en) | The virtual lift face method and apparatus that face is taken pictures | |
| CN106920277A (en) | Simulation beauty and shaping effect visualizes the method and system of online scope of freedom carving | |
| US20100189357A1 (en) | Method and device for the virtual simulation of a sequence of video images | |
| JP2019510297A (en) | Virtual try-on to the user's true human body model | |
| CN108460398B (en) | Image processing method and device and cloud processing equipment | |
| CN109801380A (en) | A kind of method, apparatus of virtual fitting, storage medium and computer equipment | |
| JP2004094917A (en) | Virtual makeup device and method therefor | |
| CN101779218A (en) | Makeup simulation system, makeup simulation apparatus, makeup simulation method, and makeup simulation program | |
| CN110363867A (en) | Virtual dress up system, method, equipment and medium | |
| CN106682632A (en) | Method and device for processing face images | |
| CN108846792A (en) | Image processing method, device, electronic equipment and computer-readable medium | |
| CN107153805A (en) | Customize makeups servicing unit and method | |
| WO2014081394A1 (en) | Method, apparatus and system for virtual clothes modelling | |
| CN114821675B (en) | Object processing method and system and processor | |
| CN107194995A (en) | A kind of method of rapid build true three-dimension person model | |
| KR20230085931A (en) | Method and system for extracting color from face images | |
| KR20230110787A (en) | Methods and systems for forming personalized 3D head and face models | |
| CN114373044A (en) | Method, device, computing equipment and storage medium for generating three-dimensional face model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |