CN110490050A - A kind of face identification method, device, system and storage medium - Google Patents
A kind of face identification method, device, system and storage medium Download PDFInfo
- Publication number
- CN110490050A CN110490050A CN201910591198.8A CN201910591198A CN110490050A CN 110490050 A CN110490050 A CN 110490050A CN 201910591198 A CN201910591198 A CN 201910591198A CN 110490050 A CN110490050 A CN 110490050A
- Authority
- CN
- China
- Prior art keywords
- face
- background
- image
- facial image
- identified
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of face identification method, device, system and storage mediums, which comprises obtains image data;Background modeling is carried out based on described image data, obtains background model;Facial image to be identified in described image data is inputted the background model progress background to filter out, obtains no background facial image;Recognition of face is carried out based on the no background facial image, obtains face recognition result.According to the method for the present invention, device, system and storage medium, pass through the background established in background model removal facial image, facial image after wiping out background is identified, eliminate the influence of background and objective environment to recognition of face, computing cost caused by and is small, and the computational efficiency, accuracy and robustness of recognition of face has been significantly greatly increased.
Description
Technical field
The present invention relates to technical field of image processing, relate more specifically to the processing of recognition of face.
Background technique
Face recognition technology is widely used to every field, with the further maturation of technology and mentioning for Social Agree
Height, face recognition technology are applied in more fields.Current face recognition technology all contains in the picture commonly entered
A certain amount of background information, these background informations can extract face characteristic and affect.Especially in side face, back
The ratio that scene area accounts for whole image is very big, causes the recognition performance of side face bad.So removal background information can promote knowledge
Not in the performance of different scenes.In general, before image partition method can be used to the background and face of image in removal background information
Scape is split, but this will cause many additional computing costs and amount of storage, cause the speed of recognition of face slow, and right
The accuracy rate of recognition of face is also without very big promotion.
Therefore, recognition of face in the prior art brings additional computing cost in the presence of segmentation background information, leads to face
The problem that the speed of identification is slow and accuracy rate is not high, influences the usage experience of user.
Summary of the invention
The present invention is proposed in view of the above problem.The present invention provides a kind of face identification method, device, system and
Computer storage medium, by establish background model remove facial image in background, to the facial image after wiping out background into
Row identification, eliminates the influence of background and objective environment to recognition of face, and caused by computing cost it is small, people has been significantly greatly increased
Computational efficiency, accuracy and the robustness of face identification.
According to the first aspect of the invention, a kind of face identification method is provided, comprising:
Obtain image data;
Background modeling is carried out based on described image data, obtains background model;
Facial image to be identified in described image data is inputted the background model progress background to filter out, obtains no back
Scape facial image;
Recognition of face is carried out based on the no background facial image, obtains face recognition result.
According to the second aspect of the invention, a kind of face identification device is provided, comprising:
Image collection module, for obtaining image data;
Background module obtains background model for carrying out background modeling based on described image data;And by described image
Facial image to be identified in data inputs the background model progress background and filters out, and obtains no background facial image;
Face recognition module obtains face recognition result for carrying out recognition of face based on the no background facial image.
According to the third aspect of the invention we, it provides a kind of face identification system, including memory, processor and is stored in
The computer program run on the memory and on the processor, which is characterized in that the processor executes the meter
The step of first aspect the method is realized when calculation machine program.
According to the fourth aspect of the invention, a kind of computer storage medium is provided, computer program is stored thereon with,
The step of being characterized in that, first aspect the method realized when the computer program is computer-executed.
Face identification method, device, system and computer storage medium according to an embodiment of the present invention, by establishing background
Model removes the background in facial image, identifies to the facial image after wiping out background, eliminates background and objective ring
Influence of the border to recognition of face, and caused by computing cost it is small, the computational efficiency, accuracy and Shandong of recognition of face has been significantly greatly increased
Stick.
Detailed description of the invention
The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention,
Feature and advantage will be apparent.Attached drawing is used to provide to further understand the embodiment of the present invention, and constitutes explanation
A part of book, is used to explain the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings,
Identical reference label typically represents same parts or step.
Fig. 1 is for realizing the signal of the exemplary electronic device of face identification method according to an embodiment of the present invention and device
Property block diagram;
Fig. 2 is the schematic flow chart of face identification method according to an embodiment of the present invention;
Fig. 3 is the schematic block diagram of face identification device according to an embodiment of the present invention;
Fig. 4 is the schematic block diagram of face identification system according to an embodiment of the present invention.
Specific embodiment
In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings
According to example embodiments of the present invention.Obviously, described embodiment is only a part of the embodiments of the present invention, rather than this hair
Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Based on described in the present invention
The embodiment of the present invention, those skilled in the art's obtained all other embodiment in the case where not making the creative labor
It should all fall under the scope of the present invention.
Firstly, being described with reference to Figure 1 the exemplary electron of the face identification method and device for realizing the embodiment of the present invention
Equipment 100.
As shown in Figure 1, electronic equipment 100 include one or more processors 101, it is one or more storage device 102, defeated
Enter device 103, output device 104, imaging sensor 105, the connection machine that these components pass through bus system 106 or other forms
The interconnection of structure (not shown).It should be noted that the component and structure of electronic equipment 100 shown in FIG. 1 are only exemplary, rather than limit
Property, as needed, the electronic equipment also can have other assemblies and structure.
The processor 101 can be central processing unit (CPU) or have data-handling capacity or instruction execution energy
The processing unit of the other forms of power, and can control other components in the electronic equipment 100 to execute desired function
Energy.
The storage device 102 may include one or more computer program products, and the computer program product can
To include various forms of computer readable storage mediums, such as volatile memory or nonvolatile memory.It is described volatile
Property memory for example may include random access memory (RAM) or cache memory (cache) etc..It is described non-volatile
Memory for example may include read-only memory (ROM), hard disk, flash memory etc..It can be on the computer readable storage medium
One or more computer program instructions are stored, processor 102 can run described program instruction, to realize sheet described below
The client functionality (realized by processor) in inventive embodiments and/or other desired functions.The computer can
It reads that various application programs and various data can also be stored in storage medium, such as the application program is used or generated various
Data etc..
The input unit 103 can be the device that user is used to input instruction, and may include keyboard, mouse, wheat
One or more of gram wind and touch screen etc..
The output device 104 can export various information (such as image or sound) to external (such as user), and
It may include one or more of display, loudspeaker etc..
Described image sensor 105 can be shot the desired image of user (such as photo, video etc.), and will be captured
Image be stored in the storage device 102 for other components use.
Illustratively, the exemplary electronic device for realizing face identification method according to an embodiment of the present invention and device can
To be implemented as smart phone, tablet computer, Image Acquisition end of access control system etc..
Face identification method 200 according to an embodiment of the present invention is described next, with reference to Fig. 2.As shown in Fig. 2, a kind of people
Face recognition method 200, comprising:
Firstly, obtaining image data in step S210;
In step S220, background modeling is carried out based on described image data, obtains background model;
In step S230, the facial image to be identified in described image data is inputted into the background model and carries out background filter
It removes, obtains no background facial image;
Finally, carrying out recognition of face in step S240 based on the no background facial image, obtaining face recognition result.
Wherein, carrying out background modeling based on described image data can be according to any image progress in image data,
Image for example including face and/or do not include face image, herein with no restrictions.
When obtaining background model using the background image occurred without face in image data, can be further improved
The accuracy of background model, can be to avoid because face or blocking for other articles cause background information not exclusively to cause background
Problem of the model to background judgement inaccuracy in image.
On the basis of establishing background model, background is carried out to the image comprising face by background model and is filtered out, only meeting
Occupy less computing resource, compared to the prior art background segment will be carried out for every facial image, and distinguish
A large amount of additional computational overheads caused by background and prospect are conducive to save computing resource, improve the calculating of subsequent recognition of face
Efficiency.In addition, can cause some influences and background that may become recognition of face objective environment such as illumination etc.
Change, background variation and objective environment such as illumination can also be eliminated by being filtered out by background model to the image progress background comprising face
Deng the influence to recognition of face, recognition of face is increased to the robustness of background variation and objective environment etc..
Illustratively, face identification method according to an embodiment of the present invention can be in setting with memory and processor
It is realized in standby, device or system.
Video face identification method according to an embodiment of the present invention can be deployed at man face image acquiring end, for example, In
Security protection application field can be deployed in the Image Acquisition end of access control system;It can also be deployed at personal terminal, such as intelligence electricity
Words, tablet computer, personal computer etc..
Alternatively, face identification method according to an embodiment of the present invention can also be deployed in server end (or cloud with being distributed
End) and personal terminal/Image Acquisition end at.For example, can go out obtain described image number at personal terminal/Image Acquisition end
According to described image data are passed to server end (or cloud) by personal terminal/Image Acquisition end, then server end (or cloud
End) background model is obtained based on described image data and carries out recognition of face.
Alternatively, face identification method according to an embodiment of the present invention can also be deployed in being distributed at Image Acquisition end and
At personal terminal.For example, can go out obtain described image data at Image Acquisition end, Image Acquisition end be by described image data
It passes at personal terminal, then obtain background model based on described image data at personal terminal and carries out recognition of face.
Face identification method according to an embodiment of the present invention, by establish background model remove facial image in background,
Facial image after wiping out background is identified, eliminates the influence of background and objective environment to recognition of face, and cause
Computing cost it is small, the computational efficiency, accuracy and robustness of recognition of face has been significantly greatly increased.
According to embodiments of the present invention, can further include in step S210:
Video image framing is carried out to the video data in described image data, generates picture frame.
In one embodiment, described image data are realtime image datas.
Wherein, image data includes video data and non-video data, and non-video data may include single-frame images, at this time
Single-frame images does not need to carry out sub-frame processing.
It should be noted that described image data are not necessarily image collecting device all picture numbers collected
According to can be only part image data therein;On the other hand, described image data can be continuous multiple image, can also
To be discontinuous, arbitrarily selected multiple image.
According to embodiments of the present invention, the method 200 further include:
The image detected in described image data whether there is face;
Using there are the described images of face in described image data as facial image to be identified, and/or by described image
Background modeling, which is carried out, there is no the described image of face in data obtains the background model.
In one embodiment, the facial image to be identified includes the facial image obtained in real time.
Illustratively, the facial image to be identified is by carrying out Face datection institute to each frame image in image data
Determining includes the picture frame of face.Specifically, such as template matching, SVM (support vector machines), neural network can be passed through
The size of the face is determined in the start image frame comprising target face etc. various method for detecting human face commonly used in the art
And position, so that it is determined that including each frame image of face in image data.It include face above by Face datection determination
The processing of picture frame be common processing in field of image processing, be no longer described in greater detail herein.
It will be appreciated that the present invention is not limited by the method for detecting human face specifically used, either existing Face datection side
Method or the in the future method for detecting human face of exploitation can be applied in face identification method according to an embodiment of the present invention, and
And it also should be including within the scope of the present invention.
According to embodiments of the present invention, can further include in step S220:
Gaussian Mixture modeling is carried out based at least part image in described image data to establish the background model;
And/or
Utilize background model described in an at least frame image update in described image data.
Wherein, the value (or feature) of each pixel is around a certain central value in the short time in certain distance in image
Interior distribution, wherein central value can be mean value, and distance can be variance.According to statistical law, if when data point is enough this
A little pixels are in Gaussian Profile, if the value of pixel deviates central value farther out, this pixel value belongs to prospect, if
The value deviation central value of pixel is close (within the scope of certain variance), then it may be said that this pixel belongs to background.Using
Described image data establish Gaussian mixture model-universal background model, and are constantly updated to the background model, can eliminate background change
Changing influences with caused by objective environment.For example, indoors in scene, since the position of image collecting device or shooting angle may
Be it is diversified, the different backgrounds in the image comprising face can be filtered out, be eliminated different by establishing background model
Background is to recognition of face bring noise, so that recognition of face is more accurate, and changes to background more robust;In addition, in room
In outer scene, since illumination changes at any time, illumination can to facial image carry out feature extraction neural network in albefaction and
Batchnorm is impacted, because having contained objective environment in the background image used when Background Modeling or update
Variation, influence of the background to reduce illumination to face can be removed by the background model that establishs or updates at this time, so that people
Face identification is more robust to illumination, further improves the accuracy of recognition of face.
Illustratively, described image data include realtime graphic or non-real-time images.That is, establishing the background mould
When type, it can be the non-real-time images data for being also possible to acquire before according to the realtime image data currently acquired.This
Outside, described image data can be according to the image data directly acquired, be also possible to what basis was obtained from other data sources.
Illustratively, at least part image includes: background image of at least part without face in described image data,
And/or at least part has the image of face.
Illustratively, Gaussian Mixture modeling is carried out to establish the back based at least part image in described image data
Scape model, comprising: the background image and/or several images for having face for obtaining several no faces establish mixed Gaussian back
Scape model is as the background model.
Illustratively, background model described in an at least frame image update in described image data is utilized, comprising:
The background model is updated using the current image frame every a preset time.
Illustratively, background model described in an at least frame image update in described image data is utilized, comprising:
Every the first preset time detection current image frame whether be the no face background image, if described current
Picture frame is the background image of the no face, then updates the background model using the current image frame;And/or
If the current image frame is not the background image of the no face, the background model is not updated.
In one embodiment, background model described in an at least frame image update in described image data is utilized, comprising:
Real-time detection current image frame whether be the no face background image;
When detecting the current image frame is the background image of the no face, the current first back without face is obtained
Scape image, and/or start to obtain current image frame as the second background image without face every the second preset time, and use
Described first background image without face and/or the second background image without face update the background model;
When detecting the current image frame, there are when face, stop updating the background model.
Wherein, since the image data of acquisition will not have always face appearance, it is understood that there may be certain time is longer
Nobody situation, at this point it is possible to be carried out every the second preset time to the background model when detecting that no face occurs
It updates, stops updating at once if having detected that face occurs.In this way, described in not only being updated always before detecting face
Background model, it is ensured that background model is consistent with real background, and is also avoided and deposited always because of timing detection
Face can not update background model and caused by computing resource waste, further reduce computing cost.
It will be appreciated that both first preset time and second preset time can be the same or different,
To be configured according to actual needs, herein with no restrictions.
Illustratively, background model described in an at least frame image update in described image data is utilized, comprising:
The background image without face in preset time period is obtained, using the back of all no faces of current slot
Background model described in scape image update.
Illustratively, background model described in an at least frame image update in described image data is utilized, comprising:
When the background image that the no face is not present in the preset time period, then the background model is not updated.
It will be appreciated that preset time period can be multiple continuous periods, it is also possible to multiple discontinuous periods, and
The length of preset time period may be the same or different, and preset time period, which can according to need, to be configured, and not limit herein
System.
In one embodiment, background model described in an at least frame image update in described image data is utilized, comprising: base
The background model is updated in following formula:
B (t+1)=B (t)+s*I, wherein B (t+1) indicates that updated model, B (t) indicate current background model, I
Indicate currently to obtain without facial image, s indicates renewal rate.
Wherein it is possible to understand, how often the renewal rate s is updated model between can indicating, i.e., with institute
Preset time period correlation is stated, s*I can indicate the background image of the no face obtained in the preset time period.
Illustratively, the method 200 further include: store the background model.In this way, more convenient for the background model
Newly;In addition, in image collecting device when leading to background model for some reason and related data is lost, it can be directly from Backup Data
The background model is obtained, computing resource is further saved.
According to embodiments of the present invention, can further include in step S230:
Each pixel in the facial image to be identified, which is calculated, based on the background model belongs to the general of background or face
Rate;
The pixel value for the pixel for belonging to background in the facial image to be identified is set to 0 based on the probability, is obtained
The no background facial image.
Wherein, background is belonged to based on each pixel in the available facial image to be identified of the background model
Probability or each pixel belong to the probability of prospect face.In order to facilitate the information extracted in image, increase the identification of computer
Efficiency can be generated according to the probability and corresponding predetermined probability threshold binarization of background parts or the pixel of face part
The masking-out of the facial image to be identified is then based on the masking-out for the pixel of background parts in the facial image to be identified
Value is set to 0, and foreground part keeps original pixel value so as to subsequent identification, to eliminate in the facial image to be identified
Background obtains no background facial image.
Illustratively, the pixel value for the pixel for belonging to background in the facial image to be identified is set based on the probability
It is 0, comprising:
The face masking-out of the facial image to be identified according to the Face image synthesis to be identified;
Based on the probability and probability threshold value by the face masking-out binaryzation, the two of the facial image to be identified are generated
Value masking-out;
Image closed operation is carried out to the binaryzation masking-out and obtains facial image masking-out to be identified;
The facial image to be identified is operated using the facial image masking-out to be identified, by the people to be identified
The pixel value for belonging to the pixel of background in face image is set to 0.
Wherein, described image closed operation can make the profile of image become smooth, can make narrow interruption and elongated up
Small hole is eliminated in gully, and face point or background parts can be made to become more continuous after image closed operation, will not
There is isolated island to influence filtering out for background parts, is conducive to the accuracy for improving removal this process of background, and then after raising
The accuracy of continuous recognition of face.
According to embodiments of the present invention, can further include in step S240:
The face characteristic model of the no background facial image input training is subjected to feature extraction and obtains face characteristic;
The face characteristic is compared with the face base map feature in the library of face bottom, obtains face recognition result.
Illustratively, the face characteristic model is obtained based on the face image data training of no background.
Due to having utilized background model to remove the background parts in facial image to be identified, so at this time to no back
Used neural network, that is, face characteristic model is also possible to based on largely without background when the progress feature extraction of scape facial image
Facial image be trained to obtain, such face characteristic model can focus more on the feature extraction of face part rather than carry on the back
The feature extraction of scape part, to improve the efficiency and accuracy of recognition of face.
Illustratively, the face characteristic is compared with the face base map feature in the library of face bottom, obtains face knowledge
Other result, comprising:
When there is the target face base map feature to match with the face characteristic in the library of the face bottom, the face
Recognition result is the target face base map feature.
Illustratively, the face characteristic is compared with the face base map feature in the library of face bottom, obtains face knowledge
Other result, further includes:
It is when the target face base map feature to match with the face characteristic is not present in the library of the face bottom, then described
Face recognition result is nothing.
In one embodiment, the target face base map feature include in the library of the face bottom with the face characteristic phase
Like the highest face base map feature of degree.
In one embodiment, when the similarity of the target face base map feature is greater than or equal to recognition threshold, institute
Stating face recognition result is the target face base map feature.
In one embodiment, the face recognition result includes the corresponding ID number of the target face base map feature.
Such as digital number 0123 indicates in the face bottom library comprising 10000 face base map features, the face that ID number is 0123
Base map feature.
In one embodiment, when the similarity of face base map feature and the face characteristic in the library of the face bottom is small
When recognition threshold, then the face recognition result is nothing.
In one embodiment, for the face identification method of the embodiment of the present invention to be deployed at Image Acquisition end,
The face identification method 200 is further illustrated.The method 200 includes:
Firstly, background model is established by camera collection image data in Image Acquisition end when initialization, it is specific to wrap
It includes: based at least part image in described image data, can be the background image of no face, establish mixed Gaussian background mould
Type stores the background model as the background model;
Then, real-time detection current image frame whether be no face background image;If the current image frame is institute
Face is not present in the background image for stating no face, then as the background image without face for updating the background model, is based on
After preset renewal rate, that is, preset time period, institute is updated using by the background image of all no faces in the preset time period
State background model;
If the background image that the current image frame is not the no face has a face, in described image data
Facial image to be identified input the background model and carry out background and filter out, obtain no background facial image, specifically include: being based on
The background model calculates the probability that each pixel in the facial image to be identified belongs to background or face;According to it is described to
Identify the face masking-out of facial image to be identified described in Face image synthesis;Based on the probability and probability threshold value by the face
Masking-out binaryzation generates the binaryzation masking-out of the facial image to be identified;Image closed operation is carried out to the binaryzation masking-out
Obtain facial image masking-out to be identified;The facial image to be identified is grasped using the facial image masking-out to be identified
Make, the pixel value for the pixel for belonging to background in the facial image to be identified is set to 0, obtains the no background face figure
Picture;
Finally, carrying out recognition of face based on the no background facial image, face recognition result is obtained, is specifically included: will
The human face recognition model of the no background facial image input training carries out feature extraction and obtains face characteristic;When the face bottom
When being greater than or equal to recognition threshold with the similarity of the highest target face base map feature of the face characteristic similarity in library, institute
Stating face recognition result is the target face base map feature;When in the library of the face bottom face base map feature and the face
When the similarity of feature is less than recognition threshold, then the face recognition result is nothing.
It follows that face identification method according to an embodiment of the present invention, removes facial image by establishing background model
In background, the facial image after wiping out background is identified, eliminates background and objective environment to the shadow of recognition of face
Ring, and caused by computing cost it is small, the computational efficiency, accuracy and robustness of recognition of face has been significantly greatly increased.
Fig. 3 shows the schematic block diagram of face identification device 300 according to an embodiment of the present invention.As shown in figure 3, according to
The face identification device 300 of the embodiment of the present invention includes:
Image collection module 310, for obtaining image data;
Background module 320 obtains background model for carrying out background modeling based on described image data;And it will be described
Facial image to be identified in image data inputs the background model progress background and filters out, and obtains no background facial image;
Face recognition module 330 obtains recognition of face knot for carrying out recognition of face based on the no background facial image
Fruit.
Wherein, face identification device 300, which carries out background modeling, can be according to any image progress in image data, example
Such as including face image and/or do not include face image, herein with no restrictions.When face identification device 300 utilizes image
The background image occurred without face in data obtains background model, it is ensured that the accuracy of background model, it will not be because of
Face or blocking for other articles cause background information not exclusively to cause background model to background judgement inaccuracy in image
Problem.On the basis of establishing background model, background is carried out to the image comprising face by background model and is filtered out, can only be occupied
Less computing resource, compared to the prior art background segment will be carried out for every facial image, and distinguish background
With a large amount of additional computational overheads caused by prospect, is conducive to save computing resource, improves the computational efficiency of subsequent recognition of face.
In addition, can cause some influences and background that may change recognition of face objective environment such as illumination etc., pass through
Background model, which filters out the image progress background comprising face, can also eliminate background variation and objective environment such as illumination etc. to people
The influence of face identification increases recognition of face to the robustness of background variation and objective environment etc..
According to embodiments of the present invention, image collection module 310 can be also used for:
Video image framing is carried out to the video data in described image data, generates picture frame.
In one embodiment, described image data are realtime image datas.
Wherein, image data includes video data and non-video data, and non-video data may include single-frame images, at this time
Single-frame images does not need to carry out sub-frame processing.
It should be noted that described image data are not necessarily image collecting device all picture numbers collected
According to can be only part image data therein;On the other hand, described image data can be continuous multiple image, can also
To be discontinuous, arbitrarily selected multiple image.
According to embodiments of the present invention, described device 300 further include:
Face detection module 340, for detecting the image in described image data with the presence or absence of face;It wherein, will be described
There are the described images of face in image data as facial image to be identified, and/or people will be not present in described image data
The described image (background image as no face) of face carries out background modeling and obtains the background model.
In one embodiment, the facial image to be identified includes the facial image obtained in real time.
Illustratively, the facial image to be identified is by carrying out Face datection institute to each frame image in image data
Determining includes the picture frame of face.Specifically, such as template matching, SVM (support vector machines), neural network can be passed through
The size of the face is determined in the start image frame comprising target face etc. various method for detecting human face commonly used in the art
And position, so that it is determined that including each frame image of face in image data.It include face above by Face datection determination
The processing of picture frame be common processing in field of image processing, be no longer described in greater detail herein.
It will be appreciated that the present invention is not limited by the method for detecting human face specifically used, either existing Face datection side
Method or the in the future method for detecting human face of exploitation can be applied in face identification method according to an embodiment of the present invention, and
And it also should be including within the scope of the present invention.
According to embodiments of the present invention, the background module 320 can be further used for:
Gaussian Mixture modeling is carried out based at least part image in described image data to establish background model;And/or
Utilize background model described in an at least frame image update in described image data.
Wherein, the value (or feature) of each pixel is to surround and a certain central value certain distance in the short time in image
Interior distribution, wherein central value can be mean value, and distance can be variance.According to statistical law, if when data point is enough this
A little points are in Gaussian Profile, if the value of pixel deviates central value farther out, this pixel value belongs to prospect, if pixel
It is close (within the scope of certain variance) that the value of point deviates central value, then it may be said that this point belongs to background.Using described image
Background image in data not comprising face establishes Gaussian mixture model-universal background model, and is constantly updated to the background model,
Influence caused by background variation and objective environment can be eliminated.Such as indoors in scene, due to the position of image collecting device
Or shooting angle may be it is diversified, the different backgrounds in the image comprising face can be filtered by establishing background model
It removes, eliminates different backgrounds to recognition of face bring noise, so that recognition of face is more accurate, and to background variation more Shandong
Stick;In addition, since illumination changes at any time, illumination can be to the nerve net for carrying out feature extraction to facial image in outdoor scene
Albefaction and batch norm in network impact, because having wrapped in the background image used when Background Modeling or update
The variation of objective environment is contained, background can have been removed at this time to reduce illumination to face by the background model establishd or updated
It influences, so that recognition of face is more robust to illumination, further improves the accuracy of recognition of face.
Illustratively, described image data include realtime graphic or non-real-time images.That is, establishing the background mould
When type, it can be the non-real-time images data for being also possible to acquire before according to the realtime image data currently acquired.This
Outside, described image data can be according to directly acquiring without image data, be also possible to what basis was obtained from other data sources.
Illustratively, at least part image includes: background image of at least part without face in described image data,
And/or at least part has the image of face.
Illustratively, the background module 320 can be further used for: obtain the background image of several no faces
And/or several images for having face, mixture Gaussian background model is established as the background model.
Illustratively, the background module 320 can be further used for: use the current figure every a preset time
As frame updates the background model.
Illustratively, the background module 320 can be further used for:
Every the first preset time detection current image frame whether be the no face background image, if described current
Picture frame is the background image of the no face, then updates the background model using the current image frame;And/or
If the current image frame is not the background image of the no face, the background model is not updated.
In one embodiment, the background module 320 can be further used for:
Real-time detection current image frame whether be the no face background image;
When detecting the current image frame is the background image of the no face, the current first back without face is obtained
Scape image, and/or start to obtain current image frame as the second background image without face every the second preset time, and use
Described first background image without face and/or the second background image without face update the background model;
When detecting the current image frame, there are when face, stop updating the background model.
Wherein, since the image data of acquisition will not have always face appearance, it is understood that there may be certain time is longer
Nobody situation, at this point it is possible to be carried out every the second preset time to the background model when detecting that no face occurs
It updates, stops updating at once if having detected that face occurs.In this way, described in not only being updated always before detecting face
Background model, it is ensured that background model is consistent with real background, and is also avoided and deposited always because of timing detection
Face can not update background model and caused by computing resource waste, further reduce computing cost.
It will be appreciated that both first preset time and second preset time can be the same or different,
To be configured according to actual needs, herein with no restrictions.
Illustratively, the background module 320 can be further used for:
The background image for obtaining the no face in preset time period, using all nothings of the current slot
The background image of face updates the background model.
Illustratively, the background module 320 can be further used for:
When the background image that the no face is not present in the preset time period, then the background model is not updated.
It will be appreciated that preset time period can be multiple continuous periods, it is also possible to multiple discontinuous periods, and
The length of preset time period may be the same or different, and preset time period, which can according to need, to be configured, and not limit herein
System.
In one embodiment, the background module 320 can be further used for: update the back based on following formula
Scape model:
B (t+1)=B (t)+s*I, wherein B (t+1) indicates that updated model, B (t) indicate current background model, I table
Show currently obtain without facial image, s indicates renewal rate.
Wherein it is possible to understand, how often the renewal rate s is updated model between can indicating, i.e., with institute
Preset time period correlation is stated, s*I can indicate the background image of the no face obtained in the preset time period.
Illustratively, described device 300 further include: memory module 350, for storing the background model.In this way, being convenient for
The update of the background model;It, can be in addition, in image collecting device when causing background model and related data to be lost for some reason
The background model is directly obtained from Backup Data, further saves computing resource.
According to embodiments of the present invention, the background module 320 can be further used for:
The probability that each pixel in the facial image to be identified belongs to face is calculated based on the background model;
The pixel value for the pixel for belonging to background in the facial image to be identified is set to 0 based on the probability, is obtained
The no background facial image.
Wherein, background is belonged to based on each pixel in the available facial image to be identified of the background model
Probability or each pixel belong to the probability of prospect face.In order to facilitate the information extracted in image, increase the identification of computer
Efficiency can be generated according to the probability and corresponding predetermined probability threshold binarization of background parts or the pixel of face part
The masking-out of the facial image to be identified is then based on the masking-out for the pixel of background parts in the facial image to be identified
Value is set to 0, and foreground part keeps original pixel value so as to subsequent identification, to eliminate in the facial image to be identified
Background obtains no background facial image.
Illustratively, the background module 320 can be further used for:
The face masking-out of the facial image to be identified according to the Face image synthesis to be identified;
Based on the probability and probability threshold value by the face masking-out binaryzation, the two of the facial image to be identified are generated
Value masking-out;
Image closed operation is carried out to the binaryzation masking-out and obtains facial image masking-out to be identified;
The facial image to be identified is operated using the facial image masking-out to be identified, by the people to be identified
The pixel value for belonging to the pixel of background in face image is set to 0.
Wherein, described image closed operation can make the profile of image become smooth, can make narrow interruption and elongated up
Small hole is eliminated in gully, and face point or background parts can be made to become more continuous after image closed operation, will not
There is isolated island to influence filtering out for background parts, is conducive to the accuracy for improving removal this process of background, and then after raising
The accuracy of continuous recognition of face.
According to embodiments of the present invention, the face recognition module 330 may include:
Characteristic extracting module 331, it is special for carrying out the face characteristic model of the no background facial image input training
Sign is extracted and obtains face characteristic;
Comparison module 332 is obtained for the face characteristic to be compared with the face base map feature in the library of face bottom
Face recognition result.
Illustratively, the face characteristic model is obtained based on the face image data training of no background.
Due to having utilized background model to remove the background parts in facial image to be identified, so at this time to no back
Used neural network, that is, face characteristic model is also possible to based on largely without background when the progress feature extraction of scape facial image
Facial image be trained to obtain, such face characteristic model can focus more on the feature extraction of face part rather than carry on the back
The feature extraction of scape part, to improve the efficiency and accuracy of recognition of face.
Illustratively, the comparison module 332 can be further used for:
When there is the target face base map feature to match with the face characteristic in the library of the face bottom, the face
Recognition result is the target face base map feature.
Illustratively, the comparison module 332 can be further used for:
It is when the target face base map feature to match with the face characteristic is not present in the library of the face bottom, then described
Face recognition result is nothing.
In one embodiment, the target face base map feature include in the library of the face bottom with the face characteristic phase
Like the highest face base map feature of degree.
In one embodiment, when the similarity of the target face base map feature is greater than or equal to recognition threshold, institute
Stating face recognition result is the target face base map feature.
In one embodiment, the face recognition result includes the corresponding ID number of the target face base map feature.
Such as digital number 0123 indicates in the face bottom library comprising 10000 face base map features, the face that ID number is 0123
Base map feature.
In one embodiment, when the similarity of face base map feature and the face characteristic in the library of the face bottom is small
When recognition threshold, then the face recognition result is nothing.
It follows that face identification device according to an embodiment of the present invention, removes facial image by establishing background model
In background, the facial image after wiping out background is identified, eliminates background and objective environment to the shadow of recognition of face
Ring, and caused by computing cost it is small, the computational efficiency, accuracy and robustness of recognition of face has been significantly greatly increased.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
Fig. 4 shows the schematic block diagram of face identification system 400 according to an embodiment of the present invention.Face identification system
400 include imaging sensor 410, storage device 420 and processor 430.
Imaging sensor 410 is for acquiring image data.
The storage of storage device 420 is for realizing the corresponding steps in face identification method according to an embodiment of the present invention
Program code.
The processor 430 is for running the program code stored in the storage device 420, to execute according to the present invention
The corresponding steps of the face identification method of embodiment, and for realizing in face identification device according to an embodiment of the present invention
Image collection module 310, background module 320 and face recognition module 330.
In addition, according to embodiments of the present invention, additionally providing a kind of storage medium, storing program on said storage
Instruction, when described program instruction is run by computer or processor for executing the face identification method of the embodiment of the present invention
Corresponding steps, and for realizing the corresponding module in face identification device according to an embodiment of the present invention.The storage medium
It such as may include the storage card of smart phone, the storage unit of tablet computer, the hard disk of personal computer, read-only memory
(ROM), Erasable Programmable Read Only Memory EPROM (EPROM), portable compact disc read-only memory (CD-ROM), USB storage,
Or any combination of above-mentioned storage medium.The computer readable storage medium can be one or more computer-readable deposit
Any combination of storage media, such as a computer readable storage medium include by being randomly generated based on action command sequence
The readable program code of calculation machine, another computer readable storage medium include for carrying out the computer-readable of recognition of face
Program code.
In one embodiment, the computer program instructions may be implemented real according to the present invention when being run by computer
Each functional module of the face identification device of example is applied, and/or recognition of face according to an embodiment of the present invention can be executed
Method.
Each module in face identification system according to an embodiment of the present invention can pass through people according to an embodiment of the present invention
The processor computer program instructions that store in memory of operation of the electronic equipment of face identification realize, or can be in root
The computer instruction stored in computer readable storage medium according to the computer program product of the embodiment of the present invention is by computer
It is realized when operation.
Face identification method, device, system and storage medium according to an embodiment of the present invention, by establishing background model
The background in facial image is removed, the facial image after wiping out background is identified, background and objective environment pair are eliminated
The influence of recognition of face, and caused by computing cost it is small, the computational efficiency, accuracy and robust of recognition of face has been significantly greatly increased
Property.
Although describing example embodiment by reference to attached drawing here, it should be understood that above example embodiment are only exemplary
, and be not intended to limit the scope of the invention to this.Those of ordinary skill in the art can carry out various changes wherein
And modification, it is made without departing from the scope of the present invention and spiritual.All such changes and modifications are intended to be included in appended claims
Within required the scope of the present invention.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, apparatus embodiments described above are merely indicative, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another equipment is closed or is desirably integrated into, or some features can be ignored or not executed.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the present invention and help to understand one or more of the various inventive aspects, In
To in the description of exemplary embodiment of the present invention, each feature of the invention be grouped together into sometimes single embodiment, figure,
Or in descriptions thereof.However, the method for the invention should not be construed to reflect an intention that i.e. claimed
The present invention claims features more more than feature expressly recited in each claim.More precisely, such as corresponding power
As sharp claim reflects, inventive point is that the spy of all features less than some disclosed single embodiment can be used
Sign is to solve corresponding technical problem.Therefore, it then follows thus claims of specific embodiment are expressly incorporated in this specific
Embodiment, wherein each, the claims themselves are regarded as separate embodiments of the invention.
It will be understood to those skilled in the art that any combination pair can be used other than mutually exclusive between feature
All features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed any method
Or all process or units of equipment are combined.Unless expressly stated otherwise, this specification (is wanted including adjoint right
Ask, make a summary and attached drawing) disclosed in each feature can be replaced with an alternative feature that provides the same, equivalent, or similar purpose.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of any
Can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize some moulds in article analytical equipment according to an embodiment of the present invention
The some or all functions of block.The present invention is also implemented as a part or complete for executing method as described herein
The program of device (for example, computer program and computer program product) in portion.It is such to realize that program of the invention can store
On a computer-readable medium, it or may be in the form of one or more signals.Such signal can be from internet
Downloading obtains on website, is perhaps provided on the carrier signal or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
The above description is merely a specific embodiment or to the explanation of specific embodiment, protection of the invention
Range is not limited thereto, and anyone skilled in the art in the technical scope disclosed by the present invention, can be easily
Expect change or replacement, should be covered by the protection scope of the present invention.Protection scope of the present invention should be with claim
Subject to protection scope.
Claims (10)
1. a kind of face identification method, which is characterized in that the described method includes:
Obtain image data;
Background modeling is carried out based on described image data, obtains background model;
Facial image to be identified in described image data is inputted the background model progress background to filter out, obtains no background people
Face image;
Recognition of face is carried out based on the no background facial image, obtains face recognition result.
2. the method as described in claim 1, which is characterized in that based on background modeling is carried out in described image data, carried on the back
Scape model, comprising:
Gaussian Mixture modeling is carried out based at least part image in described image data to establish the background model;And/or
Utilize background model described in an at least frame image update in described image data.
3. the method as described in claim 1, which is characterized in that the facial image to be identified in described image data is inputted institute
It states background model progress background to filter out, obtains no background facial image, comprising:
The probability that each pixel in the facial image to be identified belongs to background or face is calculated based on the background model;
The pixel value for the pixel for belonging to background in the facial image to be identified is set to 0 based on the probability, is obtained described
Without background facial image.
4. method as claimed in claim 3, which is characterized in that will be belonged in the facial image to be identified based on the probability
The pixel value of the pixel of background is set to 0, comprising:
The face masking-out of the facial image to be identified according to the Face image synthesis to be identified;
Based on the probability and probability threshold value by the face masking-out binaryzation, the binaryzation of the facial image to be identified is generated
Masking-out;
Image closed operation is carried out to the binaryzation masking-out and obtains facial image masking-out to be identified;
The facial image to be identified is operated using the facial image masking-out to be identified, by the face figure to be identified
The pixel value for the pixel for belonging to background as in is set to 0.
5. the method as described in claim 1, which is characterized in that carry out recognition of face based on the no background facial image, obtain
To face recognition result, comprising:
The face characteristic model of the no background facial image input training is subjected to feature extraction and obtains face characteristic;
The face characteristic is compared with the face base map feature in the library of face bottom, obtains face recognition result.
6. method as claimed in claim 5, which is characterized in that facial image number of the face characteristic model based on no background
It is obtained according to training.
7. the method as described in claim 1, which is characterized in that the method also includes:
The image detected in described image data whether there is face;
Using there are the described images of face in described image data as facial image to be identified, and/or by described image data
In there is no face described image carry out background modeling obtain the background model.
8. a kind of face identification device, which is characterized in that described device includes:
Image collection module, for obtaining image data;
Background module obtains background model for carrying out background modeling based on described image data;And by described image data
In facial image to be identified input the background model and carry out background and filter out, obtain no background facial image;
Face recognition module obtains face recognition result for carrying out recognition of face based on the no background facial image.
9. a kind of face identification system, including memory, processor and it is stored on the memory and on the processor
The computer program of operation, which is characterized in that the processor is realized in claim 1 to 7 when executing the computer program
The step of any one the method.
10. a kind of computer storage medium, is stored thereon with computer program, which is characterized in that the computer program is counted
The step of calculation machine realizes any one of claims 1 to 7 the method when executing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910591198.8A CN110490050A (en) | 2019-07-02 | 2019-07-02 | A kind of face identification method, device, system and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910591198.8A CN110490050A (en) | 2019-07-02 | 2019-07-02 | A kind of face identification method, device, system and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110490050A true CN110490050A (en) | 2019-11-22 |
Family
ID=68546648
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910591198.8A Pending CN110490050A (en) | 2019-07-02 | 2019-07-02 | A kind of face identification method, device, system and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110490050A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111914748A (en) * | 2020-07-31 | 2020-11-10 | 平安科技(深圳)有限公司 | Face recognition method and device, electronic equipment and computer readable storage medium |
CN113854981A (en) * | 2021-09-24 | 2021-12-31 | 南京六季光电技术研究院有限公司 | Millimeter wave radar-based sports multi-person vital sign monitoring method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103942539A (en) * | 2014-04-09 | 2014-07-23 | 上海交通大学 | Method for accurately and efficiently extracting human head ellipse and detecting shielded human face |
CN107292272A (en) * | 2017-06-27 | 2017-10-24 | 广东工业大学 | A kind of method and system of the recognition of face in the video of real-time Transmission |
CN108268859A (en) * | 2018-02-08 | 2018-07-10 | 南京邮电大学 | A kind of facial expression recognizing method based on deep learning |
CN108446581A (en) * | 2018-01-22 | 2018-08-24 | 北京理工雷科电子信息技术有限公司 | A kind of unmanned plane detection method in adverse circumstances |
-
2019
- 2019-07-02 CN CN201910591198.8A patent/CN110490050A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103942539A (en) * | 2014-04-09 | 2014-07-23 | 上海交通大学 | Method for accurately and efficiently extracting human head ellipse and detecting shielded human face |
CN107292272A (en) * | 2017-06-27 | 2017-10-24 | 广东工业大学 | A kind of method and system of the recognition of face in the video of real-time Transmission |
CN108446581A (en) * | 2018-01-22 | 2018-08-24 | 北京理工雷科电子信息技术有限公司 | A kind of unmanned plane detection method in adverse circumstances |
CN108268859A (en) * | 2018-02-08 | 2018-07-10 | 南京邮电大学 | A kind of facial expression recognizing method based on deep learning |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111914748A (en) * | 2020-07-31 | 2020-11-10 | 平安科技(深圳)有限公司 | Face recognition method and device, electronic equipment and computer readable storage medium |
CN111914748B (en) * | 2020-07-31 | 2023-10-27 | 平安科技(深圳)有限公司 | Face recognition method, device, electronic equipment and computer readable storage medium |
CN113854981A (en) * | 2021-09-24 | 2021-12-31 | 南京六季光电技术研究院有限公司 | Millimeter wave radar-based sports multi-person vital sign monitoring method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109117714A (en) | A kind of colleague's personal identification method, apparatus, system and computer storage medium | |
CN108876791A (en) | Image processing method, device and system and storage medium | |
CN108629791A (en) | Pedestrian tracting method and device and across camera pedestrian tracting method and device | |
CN108228421B (en) | Data monitoring method and device, computer and storage medium | |
CN108875535A (en) | image detecting method, device and system and storage medium | |
US10163043B2 (en) | System and method for facilitating logo-recognition training of a recognition model | |
CN108875540A (en) | Image processing method, device and system and storage medium | |
CN108875517A (en) | Method for processing video frequency, device and system and storage medium | |
CN106845352B (en) | Pedestrian detection method and device | |
CN106663196A (en) | Computerized prominent person recognition in videos | |
CN111008935B (en) | Face image enhancement method, device, system and storage medium | |
CN111626163B (en) | Human face living body detection method and device and computer equipment | |
CN108932456A (en) | Face identification method, device and system and storage medium | |
CN108875731A (en) | Target identification method, device, system and storage medium | |
CN109791615A (en) | For detecting and tracking the method, target object tracking equipment and computer program product of target object | |
CN110490050A (en) | A kind of face identification method, device, system and storage medium | |
CN108875509A (en) | Biopsy method, device and system and storage medium | |
CN110677585A (en) | Target detection frame output method and device, terminal and storage medium | |
CN111027450A (en) | Bank card information identification method and device, computer equipment and storage medium | |
CN108171135A (en) | Method for detecting human face, device and computer readable storage medium | |
CN111783593A (en) | Human face recognition method and device based on artificial intelligence, electronic equipment and medium | |
CN108985263A (en) | A kind of collecting method, device, electronic equipment and computer-readable medium | |
CN109829364A (en) | A kind of expression recognition method, device and recommended method, device | |
CN109508645A (en) | Personal identification method and device under monitoring scene | |
CN108875549A (en) | Image-recognizing method, device, system and computer storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191122 |