CN108932758A - Sign-in method and device based on face recognition, computer equipment and storage medium - Google Patents

Sign-in method and device based on face recognition, computer equipment and storage medium Download PDF

Info

Publication number
CN108932758A
CN108932758A CN201810713985.0A CN201810713985A CN108932758A CN 108932758 A CN108932758 A CN 108932758A CN 201810713985 A CN201810713985 A CN 201810713985A CN 108932758 A CN108932758 A CN 108932758A
Authority
CN
China
Prior art keywords
portrait
similarity
participating user
default object
arbiter
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
Application number
CN201810713985.0A
Other languages
Chinese (zh)
Inventor
陈龙
范玉龙
姚乾鹏
刘艳云
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Kingsoft Internet Security Software Co Ltd
Original Assignee
Beijing Kingsoft Internet Security Software Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Kingsoft Internet Security Software Co Ltd filed Critical Beijing Kingsoft Internet Security Software Co Ltd
Priority to CN201810713985.0A priority Critical patent/CN108932758A/en
Publication of CN108932758A publication Critical patent/CN108932758A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C1/00—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people
    • G07C1/10—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people together with the recording, indicating or registering of other data, e.g. of signs of identity
    • 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
    • G06V40/161—Detection; Localisation; Normalisation
    • 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

Abstract

The invention provides a sign-in method and device based on face recognition, computer equipment and a storage medium, wherein the method comprises the following steps: according to the participating users indicated in the activity information, sending registration information to the terminals corresponding to the participating users; the registration information is used for the terminal to access a corresponding page to input a first personal map of a participating user; acquiring a first human figure of the participating user; when the portrait is collected to obtain a second portrait, a face recognition model is adopted to recognize whether the second portrait is matched with the first portrait or not; and according to the matched first human figure, checking in the corresponding participating user. By the method, the face recognition sign-in is realized, the identity authentication of the participating user is realized while the sign-in is realized, the working efficiency is improved, and the technical problems that the traditional paper sign-in mode in the prior art is low in efficiency and the identity of the participating user cannot be automatically authenticated are solved.

Description

Method, apparatus of registering, computer equipment and storage medium based on recognition of face
Technical field
The present invention relates to technical field of mobile terminals more particularly to it is a kind of by the method, apparatus of registering of recognition of face, based on Calculate machine equipment and storage medium.
Background technique
In daily life, each enterprise, school etc. often organize various activities, such as recruitment fair, annual meeting, speech etc., And the user of activity can be organized to register.
Currently, the mode of registering mostly uses papery mode to register, after movable participating user reaches activity venue, in activity Relevant information is filled on the paper list that the side of holding provides.However, this mode efficiency of registering is lower, and ginseng can not be verified automatically With the identity information of user, it is easy to appear the case where acting as fraudulent substitute for a person.
Summary of the invention
The present invention is directed to solve at least some of the technical problems in related technologies.
For this purpose, the present invention proposes a kind of method, apparatus of registering, computer equipment and storage medium based on recognition of face, The authentication of participating user is realized to realize that recognition of face is registered, while registering, and is improved working efficiency, is solved existing skill In art traditional papery register mode low efficiency, the technical issues of identity of participating user can not be verified automatically.
The present invention proposes a kind of device of registering based on recognition of face.
The present invention proposes a kind of computer equipment.
The present invention proposes a kind of computer readable storage medium.
First aspect present invention embodiment proposes a kind of method of registering based on recognition of face, comprising:
According to the participating user indicated in action message, registration information is sent to the corresponding terminal of the participating user；Its In, the registration information, the first portrait figure for the corresponding page input participating user of the terminal access；
Obtain the first portrait figure of the participating user；
When acquisition portrait obtains the second portrait figure, using human face recognition model, the second portrait figure and first are identified Whether portrait figure matches；Wherein, the human face recognition model, for determine the second portrait figure and the first portrait figure it Between the first similarity；Determine the second similarity between the second portrait figure and the first portrait figure that remove default object, According to first similarity and second similarity, it is determined whether matching；
According to matched first portrait figure, register to corresponding participating user.
It is described to use human face recognition model as the optional implementation of another kind of first aspect present invention embodiment, Identify whether the second portrait figure and the first portrait figure match, comprising:
The generator that the second portrait figure is inputted to the human face recognition model obtains the second people for removing default object As figure；
Determine the first similarity between the second portrait figure and the first portrait figure before removing default object, and really The second similarity between fixed the second portrait figure and the first portrait figure for removing default object；
According to first similarity and second similarity, it is weighted, obtains Weighted Similarity；
Whether it is greater than similarity threshold according to the Weighted Similarity, identifies the first portrait figure and the second portrait figure is No matching.
As the optional implementation of another kind of first aspect present invention embodiment, the human face recognition model is using life At confrontation network, the human face recognition model further includes the arbiter differentiated for the output to the generator；It is described The generator that the second portrait figure is inputted to the human face recognition model, obtain removing default object the second portrait figure it Afterwards, further includes:
Using the arbiter, the second portrait figure for removing default object is differentiated, according to the differentiation The classification of device output determines whether the second portrait figure for removing default object is what the generator generated；
If the arbiter determines that the second portrait figure for removing default object is what the generator generated, to institute After stating generator adjusting parameter, the second portrait figure for removing default object is regenerated, until the arbiter is determined The second portrait figure for removing default object is not what the generator generated.
It is described using the arbiter as the optional implementation of another kind of first aspect present invention embodiment, it is right Before the second portrait figure for removing default object is differentiated, further includes:
Using the arbiter, the second portrait figure removed before presetting object is differentiated, to sentence according to The classification of other device output determines whether the second portrait figure removed before presetting object is what the generator generated；
If the arbiter determines that the second portrait figure removed before presetting object is what the generator generated, right After the arbiter adjusting parameter, the second portrait figure removed before presetting object is differentiated again, until described sentence Other device determines that the second portrait figure removed before presetting object is not what the generator generated.
As the optional implementation of another kind of first aspect present invention embodiment, before the determining default object of removal The second portrait figure and the first portrait figure between the first similarity, and according to the second portrait figure for removing default object With the second similarity between the first portrait figure, comprising:
To the second portrait figure removed before default object, the second portrait figure for removing default object and described the One portrait figure carries out key point identification, obtains key point information；
According to the key of the key point information for removing the second portrait figure before presetting object and the first portrait figure Point information, determines first similarity；
According to the key point of the key point information of the second portrait figure for removing default object and the first portrait figure Information determines second similarity.
It is described to remove default object to described as the optional implementation of another kind of first aspect present invention embodiment Preceding the second portrait figure, the second portrait figure for removing default object and the first portrait figure carries out key point identification, obtains To after key point information, further includes:
If the number for the key point that the second portrait figure removed before presetting object identifies and the removal are pre- If the number for the key point that the second portrait figure of object identifies is all larger than threshold number, determine that human image collecting angle meets limit Determine angular range.
It is described that the second portrait figure is defeated as the optional implementation of another kind of first aspect present invention embodiment The generator for entering the human face recognition model, after the second portrait figure for obtaining removing default object, further includes:
Determine that the second portrait figure removed before presetting object and the second portrait figure for removing default object are equal Imaging clearly；
And/or determine the second portrait figure removed before presetting object and second people for removing default object As in figure, human face region area accounting meets default accounting range；
And/or determine that shooting distance meets pre-determined distance range when acquiring the second portrait figure.
It is described to be indicated according in action message as the optional implementation of another kind of first aspect present invention embodiment Participating user, before sending registration information to the corresponding terminal of the participating user, further includes:
Determine not stored the first portrait figure for having the participating user in presetting database.
As the optional implementation of another kind of first aspect present invention embodiment, the action message further includes living It is dynamic to illustrate information, after whether the identification the second portrait figure and the first portrait figure match, further includes:
If matching, according to the corresponding participating user of the first portrait figure, inquires the activity declaration information, obtain described The activity prompt information that participating user is applicable in；
The activity prompt information is shown.
As the optional implementation of another kind of first aspect present invention embodiment, the method also includes:
When getting action message, the corresponding activity mark of the action message is generated；
According to the participating user indicated in the action message, the activity is sent to the corresponding terminal of the participating user Mark；
After obtaining the activity mark of participating user's input, start to acquire portrait to obtain the second portrait figure.
The method of registering based on recognition of face of the embodiment of the present invention, by the participating user indicated according to action message, The registration of the first portrait figure for the corresponding page input participating user of terminal access is sent to the corresponding terminal of participating user Information, and obtain the first portrait figure of participating user, when acquiring portrait and obtaining the second portrait figure, using human face recognition model, Identify whether the second portrait figure and the first portrait figure match, and then according to matched first portrait figure, to corresponding participating user It registers.It is matched as a result, by the second portrait figure that will be acquired with the first portrait figure of registration input, according to matched First portrait figure is registered, and is realized recognition of face and is registered, improve work efficiency, realizes while being automatically signing in ginseng With the automatic verifying of the identity information of user, it is ensured that participating user is consistent with registration user, effectively prevents acting as fraudulent substitute for a person feelings The generation of condition.
Second aspect of the present invention embodiment proposes a kind of device of registering based on recognition of face, comprising:
Sending module, for being sent out to the corresponding terminal of the participating user according to the participating user indicated in action message Send registration information；Wherein, the registration information, the first portrait for the corresponding page input parametric user of the terminal access Figure；
Module is obtained, for obtaining the first portrait figure of the participating user；
Identification module, for using human face recognition model, identifying described second when acquisition portrait obtains the second portrait figure Whether portrait figure and the first portrait figure match；Wherein, the human face recognition model, for determining the second portrait figure and described The first similarity between first portrait figure；It determines between the second portrait figure and the first portrait figure that remove default object Second similarity, according to first similarity and second similarity, it is determined whether matching；
Processing module, for being registered to corresponding participating user according to matched first portrait figure.
As the optional implementation of another kind of second aspect of the present invention embodiment, the identification module, comprising:
It is pre- to obtain removal for the second portrait figure to be inputted to the generator of the human face recognition model for generation unit If the second portrait figure of object；
Determination unit, for determining first between the second portrait figure and the first portrait figure before removing default object Similarity, and determine the second similarity between the second portrait figure and the first portrait figure that remove default object；
Computing unit, for being weighted, being weighted according to first similarity and second similarity Similarity；
Recognition unit identifies the first portrait figure for whether being greater than similarity threshold according to the Weighted Similarity Whether matched with the second portrait figure.
As the optional implementation of another kind of second aspect of the present invention embodiment, the human face recognition model is using life At confrontation network, the human face recognition model further includes the arbiter differentiated for the output to the generator；It is described Identification module, further includes:
Judgement unit, for using the arbiter, the second portrait figure for presetting object to the removal differentiates, with Determine whether the second portrait figure for removing default object is that the generator generates according to the classification of arbiter output 's；If the arbiter determines that the second portrait figure for removing default object is what the generator generated, to the life After adjusting parameter of growing up to be a useful person, regenerate the second portrait figure for removing default object, until the arbiter determine it is described The the second portrait figure for removing default object is not what the generator generated.
As the optional implementation of another kind of second aspect of the present invention embodiment, the judgement unit is also used to:
Before differentiating to the second portrait figure for removing default object, using the arbiter, gone to described Except the second portrait figure before default object is differentiated, default pair of the removal is determined with the classification exported according to the arbiter As whether the second preceding portrait figure is what the generator generated；If before the arbiter determines the default object of the removal Second portrait figure is what the generator generated, after the arbiter adjusting parameter, before presetting object to the removal again The second portrait figure differentiated, until the arbiter determines that the second portrait figure removed before default object is not institute State generator generation.
As the optional implementation of another kind of second aspect of the present invention embodiment, the determination unit is specifically used for:
To the second portrait figure removed before default object, the second portrait figure for removing default object and described the One portrait figure carries out key point identification, obtains key point information；
According to the key of the key point information for removing the second portrait figure before presetting object and the first portrait figure Point information, determines first similarity；
According to the key point of the key point information of the second portrait figure for removing default object and the first portrait figure Information determines second similarity.
As the optional implementation of another kind of second aspect of the present invention embodiment, the determination unit is also used to:
After obtaining key point information, if the key point that the second portrait figure removed before presetting object identifies Number and the number of key point that identifies of the second portrait figure for removing default object be all larger than threshold number, Determine that human image collecting angle meets restriction angular range.
As the optional implementation of another kind of second aspect of the present invention embodiment, the generation unit is also used to:
After obtaining removing the second portrait figure of default object, second portrait removed before default object is determined Figure and the second equal imaging clearly of portrait figure for removing default object；
And/or determine the second portrait figure removed before presetting object and second people for removing default object As in figure, human face region area accounting meets default accounting range；
And/or determine that shooting distance meets pre-determined distance range when acquiring the second portrait figure.
As the optional implementation of another kind of second aspect of the present invention embodiment, the sending module is also used to:
Before sending registration information to the corresponding terminal of the participating user, determine in presetting database it is not stored State the first portrait figure of participating user.
As the optional implementation of another kind of second aspect of the present invention embodiment, described device further include:
Cue module, for after identifying whether the second portrait figure and the first portrait figure match, according to described first The corresponding participating user of portrait figure inquires the activity declaration information, obtains the activity prompt information that the participating user is applicable in； The activity prompt information is shown.
As the optional implementation of another kind of second aspect of the present invention embodiment, described device further include:
Identifier generation module, for when getting action message, generating the corresponding activity mark of the action message；
The sending module, is also used to:
According to the participating user indicated in the action message, the activity is sent to the corresponding terminal of the participating user Mark；
Control module, for starting to acquire portrait to obtain after the activity mark of acquisition participating user's input To the second portrait figure.
The device of registering based on recognition of face of the embodiment of the present invention, by the participating user indicated according to action message, The registration of the first portrait figure for the corresponding page input participating user of terminal access is sent to the corresponding terminal of participating user Information, and obtain the first portrait figure of participating user, when acquiring portrait and obtaining the second portrait figure, using human face recognition model, Identify whether the second portrait figure and the first portrait figure match, and then according to matched first portrait figure, to corresponding participating user It registers.It is matched as a result, by the second portrait figure that will be acquired with the first portrait figure of registration input, according to matched First portrait figure is registered, and is realized recognition of face and is registered, improve work efficiency, realizes while being automatically signing in ginseng With the automatic verifying of the identity information of user, it is ensured that participating user is consistent with registration user, effectively prevents acting as fraudulent substitute for a person feelings The generation of condition.
Third aspect present invention embodiment proposes a kind of computer equipment, comprising: memory, processor and is stored in On reservoir and the computer program that can run on a processor, when the processor executes described program, such as aforementioned first is realized Method of registering described in aspect embodiment based on recognition of face.
Fourth aspect present invention embodiment proposes a kind of computer readable storage medium, is stored thereon with computer journey Sequence realizes the method for registering based on recognition of face as described in aforementioned first aspect embodiment when the program is executed by processor.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments Obviously and it is readily appreciated that, in which:
Fig. 1 is a kind of flow diagram of the method for registering based on recognition of face provided by the embodiment of the present invention；
Fig. 2 is the flow diagram of register method of the another kind based on recognition of face provided by the embodiment of the present invention；
Fig. 3 is the flow diagram of another method of registering based on recognition of face provided by the embodiment of the present invention；
Fig. 4 is the flow diagram of another method of registering based on recognition of face provided by the embodiment of the present invention；
Fig. 5 is a kind of structural schematic diagram of the device of registering based on recognition of face provided by the embodiment of the present invention；
Fig. 6 is the structural schematic diagram of register device of the another kind based on recognition of face provided by the embodiment of the present invention；
Fig. 7 is the structural schematic diagram of another device of registering based on recognition of face provided by the embodiment of the present invention；
Fig. 8 is the structural schematic diagram of another device of registering based on recognition of face provided by the embodiment of the present invention；With And
Fig. 9 is a kind of structural schematic diagram of computer equipment provided by the embodiment of the present invention.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
Below with reference to the accompanying drawings method, apparatus of registering, the computer equipment based on recognition of face of the embodiment of the present invention are described And storage medium.
Fig. 1 is a kind of flow diagram of the method for registering based on recognition of face provided by the embodiment of the present invention, the party Method can be applied in terminal device, also can be applied in server, wherein terminal device includes but is not limited to mobile phone, puts down The mobile electronic equipments such as plate computer, laptop.It is carried out using the method for registering provided by the invention based on recognition of face When registering, it can use terminal device in site of activity and carry out human image collecting.To be described collectively, the following embodiment of the present invention with Method of registering provided by the invention based on recognition of face carries out human image collecting using tablet computer and is applied to tablet computer Example, is described in detail the registering method proposed by the present invention based on recognition of face.
As shown in Figure 1, should based on recognition of face register method the following steps are included:
Step 101, according to the participating user indicated in action message, registration letter is sent to the corresponding terminal of participating user Breath；Wherein, registration information, the first portrait figure for the corresponding page input participating user of terminal access.
Wherein, can include but is not limited to activity name in action message, the time is held in the user of active pointer pair, activity, The information such as host place.
In daily life, most units (such as enterprise, school etc.) can all organize big and small various activities, including list Position internal activity and open activity.By taking enterprise as an example, enterprise can organize new staffs training activity, annual meeting for the employee of oneself Equal internal activities, can also organize the open activity such as special meeting, news conference.The participating user of different active pointers pair is different, For example, the new staffs training activity for business organization, only for the employee of new registration；Special meeting for tissue, only needle To meeting related personnel.
, it is understood that the activity that any activity side of holding organizes, has been intended to personnel's participation.The activity side of holding is permissible It issues and invites to desired participant, or related personnel is notified to participate in.In the present embodiment, it can use and be indicated into action message The corresponding terminal of participating user send the mode of registration information, to it is expected participant's sending activity invitation.For example, activity is lifted Registration information can be sent to activity by multiple channels such as short message, mail, instant message applications by the movable responsible person for the side of doing In the terminal of the participating user indicated in information.After participating user obtains registration information by terminal, terminal access can be passed through The corresponding page of registration information, and input in the page the first portrait figure of oneself.It wherein, include participating in using in the first portrait figure The front face at family, the first portrait figure can be the image that participating user obtains from terminal local, be also possible to lead in registration Cross the camera shooting for calling terminal.
For example, enterprise will hold primary training for the employee of new registration, and the HR (human resources) of enterprise can root The mailbox or telephone number filled according to new registration employee in registration, by way of mail or short message, to new registration employee Send the login link of training.New registration employee can carry out corresponding enrollment page by login link, in enrollment page The photo of oneself is uploaded as the first portrait figure.Alternatively, the photo that HR can also be submitted according to new registration employee in registration, from Row is for training the photo for importing new registration employee in the mobile device registered as the first portrait figure.
In a kind of possible implementation of the embodiment of the present invention, to avoid the first portrait for repeating to obtain participating user Figure, or participating user is avoided to input the first portrait figure influence user experience repeatedly, it is used in the participation indicated into action message Before the corresponding terminal in family sends registration information, not stored the first portrait for having participating user in presetting database can be first determined Figure.It is when being stored with the first portrait figure of a certain participating user in the preset database, then no longer corresponding to the participating user Terminal sends registration information；It is just corresponding to participating user in the preset database when the first portrait figure of not stored participating user Terminal send registration information so that participating user inputs the first portrait of itself by accessing the corresponding page of registration information Figure.As a result, by not stored the first portrait figure for having participating user in the preset database, just to the corresponding terminal of participating user Registration information is sent to obtain the first portrait figure of participating user, when so that participating user participating in multiple activities, on repeatedly The photo of oneself is passed, the user experience is improved.
Step 102, the first portrait figure of participating user is obtained.
In the present embodiment, after participating user inputs the first portrait figure by the corresponding page of access registration information, plate electricity The first portrait figure of the available all participation users indicated into action message of brain.
Step 103, when acquiring portrait and obtaining the second portrait figure, using human face recognition model, identify the second portrait figure and Whether the first portrait figure matches.
Wherein, human face recognition model, for determining the first similarity between the second portrait figure and the first portrait figure；It determines The second similarity between the second portrait figure and the first portrait figure of default object is removed, it is similar according to the first similarity and second Degree, it is determined whether matching.Wherein, default object can be the accessories, such as glasses, cap etc. of participating user's wearing.
After movable participating user reaches site of activity, the movable side of holding can open the camera of tablet computer to acquire The second portrait figure of participating user.After tablet computer collects the second portrait figure of participating user, recognition of face can be used Model, identifies whether the second portrait figure and the first portrait figure match.
As an example, human face recognition model can calculate the second portrait figure and each first first according to the second portrait figure Each first similarity between portrait figure generates the second people for removing default object (such as glasses) further according to the second portrait figure As figure, and then each second similarity between the second portrait figure and each first portrait figure that remove default object is determined, for every One the first portrait figure, can calculate the difference of the first similarity Yu the second similarity, and the smallest difference is corresponding the first As figure, it is determined as matching with the second portrait figure.
Step 104, it according to matched first portrait figure, registers to corresponding participating user.
It, can be according to be matched to after identifying the first portrait figure matched with the second portrait figure in the present embodiment One portrait figure, to the second portrait figure, corresponding participating user registers.
The method of registering based on recognition of face of the present embodiment, passes through the participating user indicated according to action message, Xiang Can Terminal corresponding to the user sends the registration information of the first portrait figure for the corresponding page input participating user of terminal access, And the first portrait figure of participating user is obtained, and when acquiring portrait and obtaining the second portrait figure, using human face recognition model, identification the Whether two portrait figures and the first portrait figure match, and then according to matched first portrait figure, sign to corresponding participating user It arrives.It is matched as a result, by the second portrait figure that will be acquired with the first portrait figure of registration input, according to matched the first As figure is registered, realizes recognition of face and register, improve work efficiency, realize while being automatically signing in participating user Identity information automatic verifying, it is ensured that participating user with registration user it is consistent, effectively prevent the hair for acting as fraudulent substitute for a person situation It is raw.
In order to clearly illustrate in previous embodiment, the second portrait figure and the first is identified using human face recognition model As figure whether matched specific implementation process, the method for registering the embodiment of the invention provides another kind based on recognition of face, Fig. 2 The flow diagram of method of registering for another kind provided by the embodiment of the present invention based on recognition of face.
As shown in Fig. 2, step 103 may comprise steps of on the basis of embodiment as shown in Figure 1:
Step 201, by the generator of the second portrait figure input human face recognition model, the second people for removing default object is obtained As figure.
Applicant it has been investigated that, there are certain identification error rate, this identification mistakes for current face recognition technology Rate mainly as caused by similar accessories, when carrying out recognition of face using existing face recognition technology, is easy to wear similar The different faces of accessories are identified as same people, cause to identify mistake.Wherein, accessories are, for example, glasses, cap etc..
In order to improve recognition accuracy, in the present embodiment, the second portrait figure of acquisition can be input to recognition of face mould In the generator of type, the second portrait figure for removing default object (accessories) is obtained.
As a kind of possible implementation, human face recognition model can fight network (Generative using generating Adversarial Network, GAN).It includes production from the zero-sum two-person game in game theory, GAN model that GAN, which is inspired, Model and discriminative model.Wherein, production model is used to capture the distribution of sample data, (such as high with a certain distribution is obeyed This is distributed, is uniformly distributed) noise generate the sample of a similar true training data, ultimately generate one and approach training The sample of sample.Discriminative model is two classifiers, for estimate a sample from the probability of training data, if Sample then exports maximum probability (output 1) from true training sample；If sample from the sample estimates of generation, Export small probability (output 0).
It further include the arbiter differentiated for the output to generator in the present embodiment, in human face recognition model.? Second portrait figure is input to the generator of human face recognition model, after the second portrait figure for obtaining removing default object, may be used also To differentiate to the second portrait figure for removing default object, according to arbiter using the arbiter in human face recognition model The classification of output, which determines, removes whether the second portrait figure of default object is what generator generated.If arbiter output is 0, It can determine that removing the second portrait figure of default object is what generator generated；If arbiter output is 1, it is determined that removal is pre- If the second portrait figure of object is not that generator generates.When arbiter determines that the second portrait figure for removing default object is made a living Grow up to be a useful person generation, then to generator adjusting parameter after, the second portrait figure for removing default object is regenerated, until arbiter is true Making and removing the second portrait figure of default object is not what generator generated, at this point, the default object of removal that generator generates Second portrait figure, closest to true image.As a result, by introducing arbiter, judge that generator is generated using arbiter The default object of removal the second portrait figure and true second portrait figure similarity, improve and remove the second of default object The authenticity of portrait figure, avoids image fault.
Further, in a kind of possible implementation of the embodiment of the present invention, using arbiter to the default object of removal The second portrait figure differentiated before, can also first adjust the parameter of arbiter, with guarantee arbiter differentiate accuracy.Tool Body, arbiter can be used, the second portrait figure removed before presetting object is differentiated, with point exported according to arbiter Class determines whether the second portrait figure before removing default object is what generator generated.If arbiter output is 0, judgement is gone Except the second portrait figure before default object is generated by generator；If arbiter output is 1, determine to remove default object The second preceding portrait figure is not that generator generates.The second portrait figure before arbiter determines to remove default object is to generate Device generate when, after arbiter adjusting parameter, differentiate again to removing the second portrait figure before default object, until sentencing It is not what generator generated that other device, which determines to remove the second portrait figure before default object,.When arbiter determines default pair of removal As the second preceding portrait figure be not generator generate when, it is believed that the parameter configuration of arbiter is optimal, for differentiate removal When the second portrait figure of default object, it can guarantee the accuracy for differentiating result.
In a kind of possible implementation of the embodiment of the present invention, after the second portrait figure for obtaining removing default object, Can also determine the second portrait figure before removing default object, and remove the second portrait figure of default object clarity it is high In default clarity threshold, thereby, it is possible to guarantee the picture quality of the second portrait figure for recognition of face, know to improve face Other accuracy lays the foundation；And/or determine the second portrait figure before removing default object, and remove the of default object In two portrait figures, human face region area accounting meets default accounting range, and thereby, it is possible to guarantee human face region in the second portrait figure Area accounting is suitable, avoids human face region too small the problem of causing recognition accuracy to decline；And/or determine the second people of acquisition When as figure, shooting distance meets pre-determined distance range, as a result, by acquiring the second portrait figure in suitable distance, so that acquisition The second portrait figure size it is suitable, avoid shooting distance and lead to not obtain human face region clearly the second portrait figure farther out, And then the problem of causing recognition of face to fail.
Step 202, the first similarity between the second portrait figure and the first portrait figure before removing default object is determined, with And determine the second similarity removed between the second portrait figure and the first portrait figure of presetting object.
In the present embodiment, after the second portrait figure for obtaining removing default object, can according to remove before default object the Two portrait figures, the second portrait figure and the first portrait figure for removing default object, determine second before removing default object respectively The first similarity between portrait figure and the first portrait figure, and remove default object the second portrait figure and the first portrait figure it Between the second similarity.
It, can be first pre- to removing when determining the first similarity and the second similarity as a kind of possible implementation If the second portrait figure and the first portrait figure of the default object of the second portrait figure, removal before object carry out key point identification, obtain Key point information.Wherein, key point information for example can include but is not limited to the feature of the face such as mouth, eyes, eyebrow, nose Information.In turn, the key point information of the second portrait figure before object and the key point information of the first portrait figure are preset according to removal, Determine the first similarity；Believed according to the key point of the key point information for the second portrait figure for removing default object and the first portrait figure Breath, determines the second similarity.It is closed as a result, by being identified to the second portrait figure progress key point removed before and after presetting object Key point information, and then calculate separately the first similarity and the second similarity based on key point information ensure that face and details Authenticity, it is ensured that the accuracy of similarity calculation.
Further, in a kind of possible implementation of the embodiment of the present invention, to the second people removed before presetting object As figure, the second portrait figure of the default object of removal and the progress key point identification of the first portrait figure, after obtaining key point information, also The number of the key point identified from each portrait figure can be counted.If removing the second portrait figure before default object to identify The number of the key point arrived, and remove the number of the key point that the second portrait figure of default object identifies and be all larger than threshold value Number, it is determined that human image collecting angle meets restriction angular range.By determining the second portrait figure before and after removing default object The middle number for identifying obtained key point is all larger than threshold number, it can be ensured that obtains key point as much as possible, that is, ensures to adopt The the second portrait figure integrated is laid a good foundation as face image for accurate progress recognition of face.
Step 203, it according to the first similarity and the second similarity, is weighted, obtains Weighted Similarity.
In the present embodiment, it is determined that, can be according to the first similarity and second after the first similarity and the second similarity Similarity is weighted summation, obtains Weighted Similarity.
As an example, identical weight, i.e. the first similarity can be distributed for the first similarity and the second similarity Weight with the second similarity is 0.5, that is, calculates the mean value of the first similarity and the second similarity, then Weighted Similarity =(first similarity * the second similarity of 0.5+ * 0.5).
As an example, due to default object, to recognition of face mistake, there are large effects, to guarantee recognition of face Accuracy, can the second similarity distribution between the second portrait figure and the first portrait figure to remove default object it is biggish Weight, such as the weight of the second similarity can be set to 0.6, and the weight of the first similarity is set as 0.4, then Weighted Similarity =(first similarity * the second similarity of 0.4+ * 0.6).
As an example, more similar to the first portrait figure since similarity is higher, thus, it can also be according to the first phase Respective weight is distributed like the size of degree and the second similarity, is the biggish weight of a biggish distribution in the two, for example, First similarity is greater than the second similarity, then can distribute biggish weight, such as 0.6 for the first similarity, be second similar Degree distributes lesser weight, then Weighted Similarity=(first similarity * the second similarity of 0.6+ * 0.4).
Step 204, whether similarity threshold is greater than according to Weighted Similarity, identifies the first portrait figure and the second portrait figure is No matching.
Wherein, similarity threshold is preset, such as can be set to 0.8,0.85 etc..
In the present embodiment, it is determined that after Weighted Similarity, can by Weighted Similarity and preset similarity threshold into Row compares, and then judges whether the first portrait figure and the second portrait figure match.
As an example, if only the Weighted Similarity of a first portrait figure and the second portrait figure is greater than similarity Threshold value, it is determined that the first portrait figure is matched with the second portrait figure；If there is multiple first portrait figures and the second portrait figure plus It weighs similarity and is greater than similarity threshold, then can determine the corresponding first portrait figure of maximum Weighted Similarity and the second portrait figure Matching；If being greater than similarity threshold without the Weighted Similarity of the first portrait figure and the second portrait figure, it is determined that the first portrait Figure is mismatched with the second portrait figure.
The method of registering based on recognition of face of the present embodiment, by the life that the second portrait figure is inputted to human face recognition model It grows up to be a useful person, obtains the second portrait figure for removing default object, then determine the second portrait figure and the first portrait before removing default object The first similarity between figure, and determine that the second portrait figure for removing default object is similar to second between the first portrait figure Degree, and then be weighted according to the first similarity and the second similarity, Weighted Similarity is obtained, is according to Weighted Similarity finally It is no to be greater than similarity threshold, identify whether the first portrait figure and the second portrait figure match.As a result, by before to default object is removed Similarity of the second portrait figure respectively between the first portrait figure afterwards is weighted to obtain the second portrait figure and the first portrait figure Between Weighted Similarity, judge whether the first portrait figure matches with the second portrait figure according to Weighted Similarity, can effectively keep away Exempt to misidentify problem caused by because of default object, improves the accuracy of recognition of face.
It can also include activity declaration information in action message in a kind of possible implementation of the embodiment of the present invention, When the second portrait figure for identifying participating user and the first portrait figure match, can also be said for participating user's carry out activity Bright information reminding.To which the embodiment of the present invention also proposed another method of registering based on recognition of face, and Fig. 3 is the present invention The flow diagram of another method of registering based on recognition of face provided by embodiment.
As shown in figure 3, on the basis of embodiment as shown in Figure 1, after step 103, can with the following steps are included:
Step 301, it when identifying that the second portrait figure and the first portrait figure match, is used according to the corresponding participation of the first portrait figure Family inquires activity declaration information, obtains the activity prompt information that participating user is applicable in.
Wherein, activity declaration information for example can include but is not limited to seating arrangement, present gets arrangement, order of speech Deng；The corresponding relationship of participating user and activity declaration information and activity prompt information can be stored in advance in tablet computer.
It, can be according to the first portrait figure pair when identifying that the second portrait figure and the first portrait figure match in the present embodiment The participating user answered inquires activity declaration information, obtains the activity prompt information that participating user is applicable in.
For example, participating user A is arranged at second progress New Year Message in annual meeting activity, then participate in when acquiring The second portrait figure of user A, and identify that tablet computer can be according to first with after the matched first portrait figure of the second portrait figure Portrait figure determines corresponding participating user A, and from pre-stored activity declaration information, and it is corresponding to filter out participating user A Activity declaration information, and accordingly generate the activity prompt information of participating user A.For example " participating user A makes a speech as second People carries out New Year Message ", and the activity prompt information of participating user A is obtained, for example " you will be used as second spokesman, in * * * It makes a speech later ".
Step 302, activity prompt information is shown.
In the present embodiment, after having obtained the activity prompt information that participating user is applicable in, can to activity prompt information into Row is shown, to remind participating user.
Herein it should be noted that in the embodiment of the present invention, inquires activity declaration information and show corresponding activity prompt The process of information, with the process registered according to matched first portrait figure to corresponding participating user, the two can be simultaneously It carries out, can also successively carry out, the sequencing that the present invention executes the two is with no restriction.
The method of registering based on recognition of face of the present embodiment, by identifying the second portrait figure and the matching of the first portrait figure When, according to the corresponding participating user of the first portrait figure, activity declaration information is inquired, obtains the activity prompt letter that participating user is applicable in Breath, and then activity prompt information is shown, activity schedule prompting can be carried out to participating user, participating user is facilitated to grasp Activity declaration information relevant to oneself, the user experience is improved.
On the one hand to avoid unnecessary human image collecting, the energy consumption of mobile device is saved, on the other hand at the same time for more In the case that a activity is registered, in order to enable registering more has specific aim, in a kind of possible reality of the embodiment of the present invention In existing mode, it can provide action message corresponding activity mark to participating user, only acquisition receives the participation of activity mark The second portrait figure of user.To which the embodiment of the present invention proposes another method of registering based on recognition of face, and Fig. 4 is this The flow diagram of another method of registering based on recognition of face provided by inventive embodiments.
As shown in figure 4, on the basis of embodiment as shown in Figure 1, before step 103, can with the following steps are included:
Step 401, when getting action message, the corresponding activity mark of action message is generated.
Wherein, activity mark can be by system automatically generated, and activity mark can uniquely determine corresponding activity, and can be with Activity mark after acquisition is arranged to cease to be in force automatically.
Step 402, according to the participating user indicated in action message, to the corresponding terminal transmission activity mark of participating user Know.
As an example, can be when sending registration information to the corresponding terminal of participating user, while transmission activity mark Know.
As an example, identical or different mode can be sampled, sends and infuses to the corresponding terminal of participating user after elder generation Volume information and activity mark.Registration information is sent to the corresponding terminal of participating user for example, mail can be first passed through, then by short Believe transmission activity mark.Alternatively, the first portrait figure can also be inputted by the corresponding page of access registration information in participating user It completes after registering, then is identified to the corresponding terminal transmission activity of participating user.For example, short message can be passed through on the day before activity Mode also acts while identifying to participating user's transmission activity to the corresponding terminal transmission activity mark of participating user The purpose for reminding participating user, avoids participating user from forgetting activity.
Step 403, after obtaining the activity mark of participating user's input, start to acquire portrait to obtain the second portrait figure.
In the present embodiment, after participating user reaches site of activity, the work received can be first inputted on tablet computer Dynamic mark, when tablet computer gets the activity mark of participating user's input, and activity mark and currently held movable work When dynamic mark is consistent, tablet computer just calls camera, starts acquisition portrait to obtain the second portrait figure.
The method of registering based on recognition of face of the present embodiment, by generating action message pair when getting action message The activity mark answered, and according to the participating user indicated in action message, it is identified to the corresponding terminal transmission activity of participating user, After obtaining the activity mark of participating user's input, start to acquire portrait just to obtain the second portrait figure.Pass through generation as a result, The corresponding activity mark of action message first verifies the work of participating user's input before the second portrait figure of acquisition participating user Dynamic mark starts to carry out human image collecting after being verified again, to effectively prevent using the participation indicated in non-action message Family carries out human image collecting, saves energy consumption, further improves working efficiency, it is also possible that registering with specific aim.
In order to realize above-described embodiment, the present invention also proposes a kind of device of registering based on recognition of face.
Fig. 5 is a kind of structural schematic diagram of the device of registering based on recognition of face provided by the embodiment of the present invention.
Include: sending module 510, obtain module 520, identification as shown in figure 5, being somebody's turn to do the device 50 of registering based on recognition of face Module 530 and processing module 540.
Wherein, sending module 510, for according to the participating user indicated in action message, to participating user's corresponding end End sends registration information；Wherein, registration information, the first portrait figure for the corresponding page input parametric user of terminal access.
In a kind of possible implementation of the embodiment of the present invention, sending module 510 is also used to corresponding to participating user Terminal send registration information before, determine not stored the first portrait figure for having participating user in presetting database.As a result, When participating user participates in multiple activities, without repeatedly uploading oneself photo, the user experience is improved.
Module 520 is obtained, for obtaining the first portrait figure of participating user.
Identification module 530, for using human face recognition model, identifying second when acquisition portrait obtains the second portrait figure Whether portrait figure and the first portrait figure match；Wherein, human face recognition model, for determine the second portrait figure and the first portrait figure it Between the first similarity；Determine the second similarity between the second portrait figure and the first portrait figure that remove default object, according to First similarity and the second similarity, it is determined whether matching.Wherein, default object can be the accessories of participating user's wearing, example Such as glasses, cap.
Processing module 540, for being registered to corresponding participating user according to matched first portrait figure.
In a kind of possible implementation of the embodiment of the present invention, as shown in fig. 6, on the basis of embodiment as shown in Figure 5 On, identification module 530 includes:
Generation unit 531, for obtaining removing default object by the generator of the second portrait figure input human face recognition model The second portrait figure.
In a kind of possible implementation of the embodiment of the present invention, generation unit 531 is also used to obtaining default pair of removal After the second portrait figure of elephant, the second portrait figure before removing default object is determined, and remove the second portrait of default object Equal imaging clearly is schemed, thereby, it is possible to guarantee the picture quality of the second portrait figure for recognition of face, to improve recognition of face Accuracy lays the foundation；And/or determine the second portrait figure before removing default object, and remove the second people of default object As in figure, human face region area accounting meets default accounting range, and thereby, it is possible to guarantee human face region area in the second portrait figure Accounting is suitable, avoids human face region too small the problem of causing recognition accuracy to decline；And/or determine the second portrait figure of acquisition When, shooting distance meets pre-determined distance range as a result, by acquiring the second portrait figure in suitable distance, so that the second of acquisition Portrait figure size is suitable, avoids shooting distance and leads to not obtain human face region clearly the second portrait figure farther out, and then leads The problem of causing recognition of face failure.
Determination unit 532, for determining first between the second portrait figure and the first portrait figure before removing default object Similarity, and determine the second similarity between the second portrait figure and the first portrait figure that remove default object.
Specifically, it is determined that unit 532 is used for the second portrait figure removed before presetting object, removes the second of default object Portrait figure and the first portrait figure carry out key point identification, obtain key point information；According to the second portrait removed before presetting object The key point information of the key point information of figure and the first portrait figure, determines the first similarity；The second of object is preset according to removing The key point information of the key point information of portrait figure and the first portrait figure, determines the second similarity.Thereby it is ensured that face and thin The authenticity of section, it is ensured that the accuracy of similarity calculation.
In a kind of possible implementation of the embodiment of the present invention, determination unit 532 is also used to obtaining key point information Later, if the number for the key point that the second portrait figure before removing default object identifies, and the of default object is removed The number for the key point that two portrait figures identify is all larger than threshold number, determines that human image collecting angle meets and limits angle model It encloses.By determining that the number of the key point identified in the second portrait figure before and after removing default object is all larger than threshold value Number, it can be ensured that obtain key point as much as possible, that is, ensure that the second portrait figure acquired is face image, carry out people to be accurate Face identification is laid a good foundation.
Computing unit 533, for being weighted, obtaining weighting similar according to the first similarity and the second similarity Degree.
Recognition unit 534 identifies the first portrait figure and second for whether being greater than similarity threshold according to Weighted Similarity Whether portrait figure matches.
Further, in a kind of possible implementation of the embodiment of the present invention, human face recognition model is using generation confrontation Network, human face recognition model further include the arbiter differentiated for the output to generator.To as shown in fig. 6, identification Module 530 can also include:
Judgement unit 535 differentiates the second portrait figure for removing default object, for using arbiter with basis The classification of arbiter output, which determines, removes whether the second portrait figure of default object is what generator generated；If arbiter is determined The the second portrait figure for removing default object is what generator generated, after generator adjusting parameter, regenerates default pair of removal The second portrait figure of elephant, until the second portrait figure that arbiter is determined to remove default object is not what generator generated.
To ensure that the accuracy that judgement unit 535 differentiates differentiates in a kind of possible implementation of the embodiment of the present invention Unit 535 is also used to before differentiating to the second portrait figure for removing default object, using arbiter, to default pair of removal As the second preceding portrait figure is differentiated, the second portrait figure before removing default object is determined with the classification exported according to arbiter It whether is what generator generated；If it is what generator generated that arbiter, which determines to remove the second portrait figure before default object, right After arbiter adjusting parameter, the second portrait figure removed before presetting object is differentiated again, until arbiter determines away Except the second portrait figure before default object is not what generator generated.
By adding to similarity of the second portrait figure removed before and after presetting object respectively between the first portrait figure Power obtains the Weighted Similarity between the second portrait figure and the first portrait figure, judges the first portrait figure and the according to Weighted Similarity Whether two portrait figures match, and can effectively avoid misidentifying problem caused by because of default object, improve the accuracy of recognition of face.
In a kind of possible implementation of the embodiment of the present invention, as shown in fig. 7, on the basis of embodiment as shown in Figure 5 On, it is somebody's turn to do the device 50 of registering based on recognition of face further include:
Cue module 550, for after identifying whether the second portrait figure and the first portrait figure match, according to the first portrait figure Corresponding participating user inquires activity declaration information, obtains the activity prompt information that participating user is applicable in；To activity prompt information It is shown.
By when identifying that the second portrait figure and the first portrait figure match, according to the corresponding participating user of the first portrait figure, Activity declaration information is inquired, obtains the activity prompt information that participating user is applicable in, and then be shown to activity prompt information, energy Participating user's activity schedule is enough reminded, user experience is promoted.
In a kind of possible implementation of the embodiment of the present invention, as shown in figure 8, on the basis of embodiment as shown in Figure 5 On, it is somebody's turn to do the device 50 of registering based on recognition of face further include:
Identifier generation module 500, for when getting action message, generating the corresponding activity mark of action message.
In the present embodiment, sending module 510 is also used to according to the participating user indicated in action message, to participating user couple The terminal transmission activity mark answered.
Control module 560, for starting to acquire portrait to obtain second after obtaining the activity mark of participating user's input Portrait figure.
By generating the corresponding activity mark of action message when getting action message, and indicated according in action message Participating user, to participating user's corresponding terminal transmission activity mark, after obtaining the activity mark of participating user's input, Start to acquire portrait just to obtain the second portrait figure.Effectively prevent as a result, to the participating user indicated in non-action message into Row human image collecting, saves energy consumption, further improves working efficiency, it is also possible that registering with specific aim.
It should be noted that the aforementioned explanation to the embodiment of the method for registering based on recognition of face is also applied for the reality The device of registering based on recognition of face of example is applied, details are not described herein again.
The device of registering based on recognition of face of the embodiment of the present invention, by the participating user indicated according to action message, The registration of the first portrait figure for the corresponding page input participating user of terminal access is sent to the corresponding terminal of participating user Information, and obtain the first portrait figure of participating user, when acquiring portrait and obtaining the second portrait figure, using human face recognition model, Identify whether the second portrait figure and the first portrait figure match, and then according to matched first portrait figure, to corresponding participating user It registers.It is matched as a result, by the second portrait figure that will be acquired with the first portrait figure of registration input, according to matched First portrait figure is registered, and is realized recognition of face and is registered, improve work efficiency, realizes while being automatically signing in ginseng With the automatic verifying of the identity information of user, it is ensured that participating user is consistent with registration user, effectively prevents acting as fraudulent substitute for a person feelings The generation of condition.
In order to realize above-described embodiment, the present invention also proposes a kind of computer equipment.
Fig. 9 is a kind of structural schematic diagram of computer equipment provided by the embodiment of the present invention.
As shown in figure 9, the computer equipment 60 includes: memory 601, processor 602 and is stored on memory 601 simultaneously The computer program 603 that can be run on processor 602 is realized as in the foregoing embodiment when processor 602 executes program Method of registering based on recognition of face.
In order to realize above-described embodiment, the present invention also proposes a kind of computer readable storage medium, is stored thereon with calculating Machine program realizes the method for registering based on recognition of face as in the foregoing embodiment when the program is executed by processor.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office It can be combined in any suitable manner in one or more embodiment or examples.In addition, without conflicting with each other, the skill of this field Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples It closes and combines.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or Implicitly include at least one this feature.In the description of the present invention, the meaning of " plurality " is at least two, such as two, three It is a etc., unless otherwise specifically defined.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes It is one or more for realizing custom logic function or process the step of executable instruction code module, segment or portion Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discussed suitable Sequence, including according to related function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the invention Embodiment person of ordinary skill in the field understood.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for Instruction execution system, device or equipment (such as computer based system, including the system of processor or other can be held from instruction The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set It is standby and use.For the purpose of this specification, " computer-readable medium ", which can be, any may include, stores, communicates, propagates or pass Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment It sets.The more specific example (non-exhaustive list) of computer-readable medium include the following: there is the electricity of one or more wirings Interconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory (ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable optic disk is read-only deposits Reservoir (CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other are suitable Medium, because can then be edited, be interpreted or when necessary with it for example by carrying out optical scanner to paper or other media His suitable method is handled electronically to obtain described program, is then stored in computer storage.
It should be appreciated that each section of the invention can be realized with hardware, software, firmware or their combination.Above-mentioned In embodiment, software that multiple steps or method can be executed in memory and by suitable instruction execution system with storage Or firmware is realized.Such as, if realized with hardware in another embodiment, following skill well known in the art can be used Any one of art or their combination are realized: have for data-signal is realized the logic gates of logic function from Logic circuit is dissipated, the specific integrated circuit with suitable combinational logic gate circuit, programmable gate array (PGA), scene can compile Journey gate array (FPGA) etc..
Those skilled in the art are understood that realize all or part of step that above-described embodiment method carries It suddenly is that relevant hardware can be instructed to complete by program, the program can store in a kind of computer-readable storage medium In matter, which when being executed, includes the steps that one or a combination set of embodiment of the method.
It, can also be in addition, each functional unit in each embodiment of the present invention can integrate in a processing module It is that each unit physically exists alone, can also be integrated in two or more units in a module.Above-mentioned integrated mould Block both can take the form of hardware realization, can also be realized in the form of software function module.The integrated module is such as Fruit is realized and when sold or used as an independent product in the form of software function module, also can store in a computer In read/write memory medium.
Storage medium mentioned above can be read-only memory, disk or CD etc..Although having been shown and retouching above The embodiment of the present invention is stated, it is to be understood that above-described embodiment is exemplary, and should not be understood as to limit of the invention System, those skilled in the art can be changed above-described embodiment, modify, replace and become within the scope of the invention Type.

Claims (10)

1. a kind of method of registering based on recognition of face, which is characterized in that the described method comprises the following steps:
According to the participating user indicated in action message, registration information is sent to the corresponding terminal of the participating user；Wherein, institute Registration information is stated, the first portrait figure for the corresponding page input participating user of the terminal access；
Obtain the first portrait figure of the participating user；
When acquisition portrait obtains the second portrait figure, using human face recognition model, the second portrait figure and the first portrait are identified Whether figure matches；Wherein, the human face recognition model, for determining between the second portrait figure and the first portrait figure First similarity；Determine the second similarity between the second portrait figure and the first portrait figure that remove default object, according to First similarity and second similarity, it is determined whether matching；
According to matched first portrait figure, register to corresponding participating user.
2. method according to claim 1 of registering, which is characterized in that it is described to use human face recognition model, identify described the Whether two portrait figures and the first portrait figure match, comprising:
The generator that the second portrait figure is inputted to the human face recognition model obtains the second portrait for removing default object Figure；
Determine that the first similarity between the second portrait figure and the first portrait figure before removing default object, and determination are gone Except the second similarity between the second portrait figure and the first portrait figure of default object；
According to first similarity and second similarity, it is weighted, obtains Weighted Similarity；
Whether be greater than similarity threshold according to the Weighted Similarity, identify the first portrait figure and the second portrait figure whether Match.
3. method according to claim 2 of registering, which is characterized in that the human face recognition model is using generation confrontation net Network, the human face recognition model further include the arbiter differentiated for the output to the generator；It is described by described Two portrait figures input the generator of the human face recognition model, after the second portrait figure for obtaining removing default object, further includes:
Using the arbiter, the second portrait figure for removing default object is differentiated, with defeated according to the arbiter Classification out determines whether the second portrait figure for removing default object is what the generator generated；
If the arbiter determines that the second portrait figure for removing default object is what the generator generated, to the life After adjusting parameter of growing up to be a useful person, regenerate the second portrait figure for removing default object, until the arbiter determine it is described The the second portrait figure for removing default object is not what the generator generated.
4. method according to claim 3 of registering, which is characterized in that it is described to use the arbiter, it is pre- to the removal Before if the second portrait figure of object is differentiated, further includes:
Using the arbiter, the second portrait figure removed before presetting object is differentiated, according to the arbiter The classification of output determines whether the second portrait figure removed before presetting object is what the generator generated；
If the arbiter determines that the second portrait figure removed before presetting object is what the generator generated, to described After arbiter adjusting parameter, the second portrait figure removed before presetting object is differentiated again, until the arbiter Determine that the second portrait figure removed before presetting object is not what the generator generated.
5. method according to claim 2 of registering, which is characterized in that determining second portrait removed before default object The first similarity between figure and the first portrait figure, and according to the second portrait figure and described first for removing default object The second similarity between portrait figure, comprising:
To the second portrait figure removed before default object, the second portrait figure for removing default object and described the first As figure progress key point identification, key point information is obtained；
Believed according to the key point information for removing the second portrait figure before presetting object and the key point of the first portrait figure Breath, determines first similarity；
The key point information of the second portrait figure of object and the key point information of the first portrait figure are preset according to the removal, Determine second similarity.
6. preordering method according to claim 5, which is characterized in that described to second people removed before presetting object As figure, the second portrait figure for removing default object and the first portrait figure progress key point identification, key point letter is obtained After breath, further includes:
If the number for the key point that the second portrait figure removed before presetting object identifies and default pair of the removal The number for the key point that the second portrait figure of elephant identifies is all larger than threshold number, determines that human image collecting angle meets and limits angle Spend range.
7. method according to claim 2 of registering, which is characterized in that described that the second portrait figure is inputted the face The generator of identification model, after the second portrait figure for obtaining removing default object, further includes:
Determine the clear of the second portrait figure of the second portrait figure removed before presetting object and the default object of the removal Degree is above default clarity threshold；
And/or determine the second portrait figure and the second portrait figure for removing default object removed before default object In, human face region area accounting meets default accounting range；
And/or determine that shooting distance meets pre-determined distance range when acquiring the second portrait figure.
8. method according to claim 1-7 of registering, which is characterized in that described according to being indicated in action message Participating user, before sending registration information to the corresponding terminal of the participating user, further includes:
Determine not stored the first portrait figure for having the participating user in presetting database.
9. method according to claim 1-7 of registering, which is characterized in that the action message further includes activity Illustrate information, after whether the identification the second portrait figure and the first portrait figure match, further includes:
If matching, according to the corresponding participating user of the first portrait figure, inquires the activity declaration information, obtains the participation The activity prompt information that user is applicable in；
The activity prompt information is shown.
10. a kind of device of registering based on recognition of face, which is characterized in that described device includes:
Sending module, for sending note to the corresponding terminal of the participating user according to the participating user indicated in action message Volume information；Wherein, the registration information, the first portrait figure for the corresponding page input parametric user of the terminal access；
Module is obtained, for obtaining the first portrait figure of the participating user；
Identification module, for using human face recognition model, identifying second portrait when acquisition portrait obtains the second portrait figure Whether figure and the first portrait figure match；Wherein, the human face recognition model, for determining the second portrait figure and described first The first similarity between portrait figure；Determine second between the second portrait figure and the first portrait figure that remove default object Similarity, according to first similarity and second similarity, it is determined whether matching；
Processing module, for being registered to corresponding participating user according to matched first portrait figure.
CN201810713985.0A 2018-06-29 2018-06-29 Sign-in method and device based on face recognition, computer equipment and storage medium Pending CN108932758A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810713985.0A CN108932758A (en) 2018-06-29 2018-06-29 Sign-in method and device based on face recognition, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810713985.0A CN108932758A (en) 2018-06-29 2018-06-29 Sign-in method and device based on face recognition, computer equipment and storage medium

Publications (1)

Publication Number Publication Date
CN108932758A true CN108932758A (en) 2018-12-04

Family

ID=64447681

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810713985.0A Pending CN108932758A (en) 2018-06-29 2018-06-29 Sign-in method and device based on face recognition, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN108932758A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112634483A (en) * 2020-12-24 2021-04-09 深兰盛视科技（苏州）有限公司 Sign-in method, device, electronic equipment and storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2653772A1 (en) * 2012-04-20 2013-10-23 Honeywell International Inc. Image recognition for personal protective equipment compliance enforcement in work areas
CN104506562A (en) * 2015-01-13 2015-04-08 东北大学 Two-dimension code and face recognition fused conference identity authentication device and method
CN105469076A (en) * 2015-12-29 2016-04-06 东莞市微模式软件有限公司 Face comparing verification method based on multi-instance learning
CN107506708A (en) * 2017-08-14 2017-12-22 广东欧珀移动通信有限公司 Solve lock control method and Related product
CN107729822A (en) * 2017-09-27 2018-02-23 北京小米移动软件有限公司 Object identifying method and device
CN107945118A (en) * 2017-10-30 2018-04-20 南京邮电大学 A kind of facial image restorative procedure based on production confrontation network

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2653772A1 (en) * 2012-04-20 2013-10-23 Honeywell International Inc. Image recognition for personal protective equipment compliance enforcement in work areas
CN104506562A (en) * 2015-01-13 2015-04-08 东北大学 Two-dimension code and face recognition fused conference identity authentication device and method
CN105469076A (en) * 2015-12-29 2016-04-06 东莞市微模式软件有限公司 Face comparing verification method based on multi-instance learning
CN107506708A (en) * 2017-08-14 2017-12-22 广东欧珀移动通信有限公司 Solve lock control method and Related product
CN107729822A (en) * 2017-09-27 2018-02-23 北京小米移动软件有限公司 Object identifying method and device
CN107945118A (en) * 2017-10-30 2018-04-20 南京邮电大学 A kind of facial image restorative procedure based on production confrontation network

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112634483A (en) * 2020-12-24 2021-04-09 深兰盛视科技（苏州）有限公司 Sign-in method, device, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
CN108875989A (en) Reservation method and device based on face recognition, computer equipment and storage medium
CN105303632B (en) A kind of mobile monitor is registered system and method for work
CN105719357B (en) Recognition of face class hour calculation method, apparatus and system
CN109120605A (en) Authentication and account information variation and device
US9065787B2 (en) Method for supervising a chat connection
CN110210194A (en) Electronic contract display methods, device, electronic equipment and storage medium
CN109858222B (en) Identity verification block chain system based on human face features and verification method thereof
CN208351494U (en) Face identification system
Naen et al. Development of attendance monitoring system with artificial intelligence optimization in cloud
CN108932758A (en) Sign-in method and device based on face recognition, computer equipment and storage medium
Shafi et al. Fingerprint verification over the network and its application in attendance management
Rufai et al. A biometric model for examination screening and attendance monitoring in Yaba College of Technology
Sayed et al. Biometrics: effectiveness and applications within the blended learning environment
CN201522734U (en) Iris recognition entrance guard
Varma et al. Aadhar card verification base online polling
Rahmatya et al. Design of student attendance information system with fingerprints
CN108764149A (en) A kind of training method for class student faceform
CN109344378A (en) A kind of smart pen record manufacturing system
CN108710820A (en) Infantile state recognition methods, device and server based on recognition of face
CN111192150B (en) Method, device, equipment and storage medium for processing vehicle danger-giving agent service
EP1784764A1 (en) Biometric identification system
CN111695571A (en) Notarization file verification method and device, computer equipment and storage medium
CN113450121B (en) Face recognition method for campus payment
CN110443122A (en) Information processing method and Related product
Saini et al. Comparative Analysis of Various Biometric Techniques for Database Security

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: 20181204