CN106326857A - Gender identification method and gender identification device based on face image - Google Patents
Gender identification method and gender identification device based on face image Download PDFInfo
- Publication number
- CN106326857A CN106326857A CN201610698494.4A CN201610698494A CN106326857A CN 106326857 A CN106326857 A CN 106326857A CN 201610698494 A CN201610698494 A CN 201610698494A CN 106326857 A CN106326857 A CN 106326857A
- Authority
- CN
- China
- Prior art keywords
- facial image
- image
- target
- sex
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- 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
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/247—Aligning, centring, orientation detection or correction of the image by affine transforms, e.g. correction due to perspective effects; Quadrilaterals, e.g. trapezoids
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Probability & Statistics with Applications (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a gender identification method and a gender identification device based on a face image. The gender identification method comprises the steps of through a prior model in a preset deep convolutional neural network, performing gender classification determining on an objective face image, and obtaining a gender determining result which corresponds with the objective face image; through performing comparison for obtaining cosine similarity between the characteristic of the objective face image and the characteristic of an image sample, obtaining a cosine comparing result which corresponds with the objective face image; performing Bayes classifier operation on the gender determining result and the cosine comparing result, obtaining the gender of an object in the face image. Therefore, in a complicated imaging condition, through splicing key parts in the face image, and combining deep convolutional neural network determining and image characteristic comparing, auxiliary identification can be performed by means of the characteristic of each key part area, deep texture, edge and color characteristic of the face, thereby improving gender identification accuracy.
Description
Technical field
The invention belongs to image identification technical field, particularly relate to a kind of gender identification method based on facial image and dress
Put.
Background technology
Recognition of face, is facial feature information based on people a kind of biological identification technology of carrying out identification.With shooting
Machine or camera collection contain image or the video flowing of face, and detect and track face the most in the picture, are generally also called
Identification of Images, facial recognition.Face recognition products be widely used to finance, the administration of justice, army, public security, frontier inspection, government, space flight,
The fields such as electric power, factory, education, medical and numerous enterprises and institutions.Such as, recognition of face access control and attendance system and recognition of face
Antitheft door, about computer login, E-Government and the ecommerce of information security.The most ripe and social along with technology
The raising of degree of recognition, face recognition technology is applied in more field.Wherein, sex identification based on image can be effectively
Auxiliary recognition of face, the accuracy of sex identification can directly affect the accuracy of final recognition of face.
In prior art, the flow process of facial image sex identification is: advanced row Face datection, the then feature of facial image
Extract, finally according to the feature the extracted sex by grader identification facial image.Due to the relation of imaging device, image is not
Being bound to react clearly the full facial of people, most image there will be unsharp situations such as fuzzy, high light, dark,
Or in image, bowing occurs in personage, side face etc. can not show the attitude of whole face completely, at the image-forming condition of this complexity
Under, the feature extracting facial image that prior art is simple is the sex that cannot accurately differentiate personage, and then can increase identification
The error rate of result.
Summary of the invention
The embodiment of the present invention provides a kind of gender identification method based on facial image and device, it is intended to solve due to outside
The change of imaging factors and the feature that cannot extract facial image accurately that causes, so can increase the asking of error rate of identification
Topic.
A kind of based on facial image the gender identification method that the embodiment of the present invention provides, including: by the preset degree of depth
Prior model in convolutional neural networks, carries out sex discriminant classification to target facial image, obtains described target facial image pair
The Sex Discrimination result answered, the image that wherein said target facial image is made up of key position region each in facial image；
By the cosine similarity between feature and the feature of image pattern of target facial image described in comparison, obtain described target person
The cosine comparison result that face image is corresponding；Described Sex Discrimination result and described cosine comparison result are carried out Bayes classifier
Computing, obtains the sex of object in described facial image.
A kind of based on facial image the sex identification device that the embodiment of the present invention provides, including: Sex Discrimination module is used
Prior model in by preset degree of depth convolutional neural networks, carries out sex discriminant classification to target facial image, obtains institute
Stating the Sex Discrimination result that target facial image is corresponding, wherein said target facial image is by key position each in facial image
The image of region composition；Comparison processing module is for by the feature of the feature of target facial image described in comparison Yu image pattern
Between cosine similarity, obtain the cosine comparison result that described target facial image is corresponding；Identification module is for by described property
Result and described cosine comparison result Pan Bie not carry out Bayes classifier computing, obtain the property of object in described facial image
Not.
The gender identification method based on facial image of embodiment of the present invention offer and device, by preset degree of depth convolution
Prior model in neutral net, carries out sex discriminant classification to target facial image, obtains described target facial image corresponding
Sex Discrimination result, the image that wherein said target facial image is made up of key position region each in facial image, passes through
Cosine similarity between feature and the feature of image pattern of target facial image described in comparison, obtains described target face figure
As corresponding cosine comparison result, described Sex Discrimination result and described cosine comparison result are carried out Bayes classifier fortune
Calculate, obtain the sex of object in described facial image, so under complicated imaging contexts, by key each in facial image
Position re-starts splicing, and degree of depth convolutional neural networks differentiating, the aspect ratio with image is to combining, it is possible to utilize face
The feature in each key position region, depth texture, edge assist in identifying with color characteristic, add the accurate of sex identification
Property.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is only this
Some embodiments of invention.
Fig. 1 is that the flow process that realizes of the gender identification method based on facial image that first embodiment of the invention provides is illustrated
Figure；
Fig. 2 is that the flow process that realizes of the gender identification method based on facial image that second embodiment of the invention provides is illustrated
Figure；
Fig. 3 is the schematic diagram of Face datection in the embodiment of the present invention；
Fig. 4 is the schematic diagram of facial image in the embodiment of the present invention；
Fig. 5 is the schematic diagram of the target facial image obtained by each key position regional restructuring in the embodiment of the present invention；
Fig. 6 is the structural representation of the sex identification device based on facial image that third embodiment of the invention provides；
Fig. 7 is the structural representation of the sex identification device based on facial image that fourth embodiment of the invention provides.
Fig. 8 is the hardware of the electronic equipment of the gender identification method based on facial image that fifth embodiment of the invention provides
Structural representation.
Detailed description of the invention
For making the goal of the invention of the present invention, feature, the advantage can be the most obvious and understandable, below in conjunction with the present invention
Accompanying drawing in embodiment, is clearly and completely described the technical scheme in the embodiment of the present invention, it is clear that described reality
Executing example is only a part of embodiment of the present invention, and not all embodiments.Based on the embodiment in the present invention, people in the art
The every other embodiment that member is obtained under not making creative work premise, broadly falls into the scope of protection of the invention.
Refer to the realization of the gender identification method based on facial image that Fig. 1, Fig. 1 provide for first embodiment of the invention
Schematic flow sheet, can be applicable in the terminal that face Time Attendance Device, face burglary-resisting system, computer etc. identify facial image.Fig. 1
Shown gender identification method based on facial image, mainly comprises the steps that
S101, by prior model in preset degree of depth convolutional neural networks, target facial image is carried out Gender Classification
Differentiate, obtain the Sex Discrimination result that this target facial image is corresponding.
The image that this target facial image is made up of key position region each in facial image.This each key position region
Can be region and the face contour region of each organ sites of face, herein to target facial image Zhong Ge key position district
Territory does not limits.It should be noted that the relation of facial image and target person image is: this target facial image is by this face
The image of this each key position region composition in image.This facial image is for representing the image of overall face.
Degree of depth convolutional neural networks is convolutional neural networks (CNN, Convolutional Neural Networks), is
A kind of deep neural network with convolutional coding structure, at least include two non-linear trainable convolutional layers, two nonlinear
Fixing convolutional layer and full articulamentum, at least 5 hidden layers, are mainly used in speech analysis and field of image recognition altogether.
Can be obtained by prior model by degree of depth convolutional neural networks training image sample, wherein training image sample with
The process obtaining prior model is the process classified image pattern according to preset classification, and wherein prior model can be used for
Image is carried out the model of discriminant classification.
S102, by the cosine similarity between feature and the feature of image pattern of this target facial image of comparison,
To the cosine comparison result that this target facial image is corresponding.
In field of image recognition, the feature of image is the proper noun of field of image recognition, the extraction of the feature of image
It it is a concept in computer vision and image procossing.
The feature of image pattern can be stored in the memory module of terminal built-in, it is also possible to is stored in cloud server
On.Cosine similarity is as the size weighing two interindividual variations by two vectorial angle cosine values in vector space
Tolerance.
S103, this Sex Discrimination result and this cosine comparison result are carried out Bayes classifier computing, obtain this face
The sex of objects in images.
It should be noted that object is the people described in this facial image in this facial image.Bayes classifier is
Calculating object by Bayesian formula and belong to the algorithm of which class probability, the principle of Bayes classifier computing is by certain
The prior probability of object, utilizes Bayesian formula to calculate the posterior probability of this object, i.e. this object and belongs to the probability of a certain class,
Select there is the classification of maximum a posteriori probability as the classification belonging to this object.
Using this Sex Discrimination result and this cosine comparison result as object in facial image in the embodiment of the present invention
Prior probability, calculates the male's classification belonging to this object and the posterior probability of women classification by Bayesian formula, and selects
Take the sex classification with maximum a posteriori probability as the sex classification of this object.
In the embodiment of the present invention, by prior model in preset degree of depth convolutional neural networks, target facial image is entered
Row Gender Classification differentiates, obtaining the Sex Discrimination result that this target facial image is corresponding, wherein this target facial image is by people
The image of each key position region composition in face image, then by the feature of this target facial image of comparison and image pattern
Cosine similarity between feature, obtains the cosine comparison result that this target facial image is corresponding, is finally tied by this Sex Discrimination
Fruit and this cosine comparison result carry out Bayes classifier computing, obtain the sex of object in this facial image, so in complexity
Imaging contexts under, by key position each in facial image being re-started splicing, and by degree of depth convolutional neural networks differentiate
With the aspect ratio of image to combining, it is possible to utilize the feature in face each key position region, depth texture, edge special with color
Levy and assist in identifying, add the accuracy of sex identification.
Refer to the realization of the gender identification method based on facial image that Fig. 2, Fig. 2 provide for second embodiment of the invention
Schematic flow sheet, can be applicable in the terminal that face Time Attendance Device, face burglary-resisting system, computer etc. identify facial image, main
Comprise the following steps:
S201, according to position in this facial image of each key point of facial image, extract in this facial image crucial
Area.
S202, according to the image of key position regional restructuring face in this facial image, obtain this target facial image.
The image that this target facial image is made up of key position region each in facial image.
Alternatively, also included before step S201: determine to be identified by Face datection and face key point location
Facial image in image；It is set as the facial image in this image to be identified detecting region.
By small echo (HAAR) grader or DLIB (C++library) algorithm, the image of input is carried out Face datection,
Then the image after detection is carried out face pass by supervised descent algorithm (SDM, Supervised Descent Method)
Key point location, is wherein included by the face key point that SDM algorithm positions: eyebrow, eyes, nose, face and face contour.When
So Face datection and face key point location can also be realized by other algorithms.
HAAR grader, comprises self adaptation and strengthens (Adaboost) algorithm, and in field of image recognition, grader refers to right
Face and the non-face algorithm carrying out classifying.DLIB is the algorithms library of a kind of C++, can be applicable to Face datection and face key point
Location.
Fig. 3 is the schematic diagram of Face datection, as it is shown on figure 3, atrous square box is face detection block, and circular expression people
Face, triangular representation animal, Polygons Representation trees, facial image can be extracted in the picture through Face datection, and by people
Face image is arranged in Face datection frame.
Alternatively, expand this detection region according to preset times, so that this facial image includes the image of hair zones.
Expanding this detection region according to preset times is to extend in facial image by hair zones, and the mode expanding detection region is permissible
It is that whole detection region is enlarged, it is also possible to be that the detection top in region, bottom, left and right two parts are all expanded preset times
Number, it is also possible to only expand the upper and lower in this detection region.This preset times comes directly towards to send out with hair lengths and brain highly to be had
Closing, in this enforcement, the numerical value of preferred preset times is 0.15, the upper and lower in this detection region will respectively expand 0.15 times.
Fig. 4 is the schematic diagram of facial image, is wherein detection region within rectangle frame in Fig. 4 (a) and Fig. 4 (b), the inspection in Fig. 4 (a)
Survey in region and include expanding the facial image before detection region, include in the detection region in Fig. 4 (b) expanding detection region it
After facial image.
Owing to color, length and the hair style of hair are to identify the foundation that sex is critically important, so facial image includes
Hair zones can more accurately identify sex.
Alternatively, eye areas in this detection region is corrected, so that in this facial image in this eye areas
Eyes are on same level line.The mode of correction ocular does not limits, and can carry out correction ocular by geometric transformation, it is also possible to
Changing angle between two eyes and carry out correction ocular, final purpose is that the eyes in this eye areas are in same level line
On.
Face key point includes: eyes, nose, face, eyebrow, hair and face contour, by each key point at this
Position in facial image, extracts key position region in this facial image.Wherein key position region bag in this facial image
Include: hair zones, brow region, eye areas, nasal area, face region.By reconstructing what each key position region obtained
Target facial image can describe the office of the faces such as the hair style of face, skin and the face contour of different sexes more accurately
Region, portion.
As it is shown in figure 5, the schematic diagram that Fig. 5 is the target facial image obtained by each key position regional restructuring.In order to just
In explanation and display, in the schematic diagram shown in Fig. 5 key position region only include eye areas, brow region, nasal area and
Face region, in Figure 5, the schematic diagram shown in Fig. 5 is only a citing to the not display of other key position regions, not
Key position region in the present invention can be constituted and limit.
S203, by prior model in preset degree of depth convolutional neural networks, target facial image is carried out Gender Classification
Differentiate, obtain the Sex Discrimination result that this target facial image is corresponding.
Degree of depth convolutional neural networks is a kind of deep neural network with convolutional coding structure, at least include two non-linear can
The convolutional layer of training, two nonlinear fixing convolutional layers and full articulamentum, at least 5 hidden layers, are mainly used in language altogether
Cent analysis and field of image recognition.Explanation to convolutional neural networks, prior model herein refer to the present invention first
Associated description in embodiment, does not repeats.
Alternatively, by prior model in preset degree of depth convolutional neural networks, target facial image is carried out sex and divides
Class differentiates, obtains the Sex Discrimination result that this target facial image is corresponding, particularly as follows:
By prior model in this degree of depth convolutional neural networks, this target facial image is carried out discriminant classification, is somebody's turn to do
Most probable value that target facial image is corresponding and first object sex classification corresponding to this most probable value；
Using this most probable value and this first object sex classification as described Sex Discrimination result.
This sex classification includes women classification and male's classification.To multiple key position figures in degree of depth convolutional neural networks
Decent is trained, and obtains prior model after training.Be can be obtained by by degree of depth convolutional neural networks training image sample
Prior model, wherein training image sample is to enter image pattern according to preset sex classification to obtain the process of prior model
The process of row classification, wherein prior model can be used for carrying out image the model of discriminant classification.This first object sex classification is
Male's classification or women classification.
S204, by the cosine similarity between feature and the feature of image pattern of this target facial image of comparison,
To the cosine comparison result that this target facial image is corresponding.
This image pattern is the image pattern being made up of each key position region in overall face image pattern, and wherein this is whole
In body facial image sample, key position region includes: hair zones, brow region, eye areas, nasal area, face district
Territory.The region that in this entirety face image pattern, key position region and key position region in this facial image include is right
Answer, i.e. if this entirety face image pattern includes brow region, then this facial image includes brow region, if this is whole
Body facial image sample includes lens area, then this facial image includes lens area.The image pattern obtained after reconstruct
Composition and the image similarity shown in Fig. 5.The mode in key position region is extracted and from face from overall face image pattern
The extracting mode in image zooming-out key position region is identical, does not repeats.Same, reconstruct obtains the reconstruct of image pattern
The reconstruct mode that mode obtains target facial image with reconstruct is identical, does not repeats.
This cosine similarity is as weighing two interindividual variations by two vectorial angle cosine values in vector space
The tolerance of size.
Alternatively, by the cosine similarity between feature and the feature of image pattern of this target facial image of comparison,
Obtain the cosine comparison result that this target facial image is corresponding, particularly as follows:
By by carrying out cosine similarity comparison between feature and the feature of image pattern of this target facial image, obtaining
Multiple cosine similarity values that this target facial image is corresponding；
According to this cosine similarity value order from high to low, choose the target face image pattern of preset number；
The sex classification of this target face image pattern is carried out quantity statistics, and using sex classifications most for quantity as
Second target gender classification；
The similar face image that sex classification is this second target gender classification is chosen in this target face image pattern
Sample, and extract the cosine similarity value between this similar face image sample with this target facial image as target cosine phase
Like angle value；
Calculate the meansigma methods of this target cosine similarity value；
Using this second target gender classification and described meansigma methods as this cosine comparison result.
The numerical value of cosine similarity value is the biggest, represents that similarity is the highest.First according to the cosine similarity value calculated, will figure
Decent, arranged by height order on earth according to similarity, then by similarity height, select the mesh of preset number
Mark facial image sample.Preset number is arbitrarily chosen, and the number certainly chosen is many, then the last accuracy differentiated will improve.
The preferred preset number of the present embodiment is 20.
The sex classification of this target face image pattern is carried out quantity statistics, and by sex classifications most for sample size
As target gender classification, if the quantity of the sample of male's classification is many in this target face image pattern, then target gender class
Wei male；If the quantity of the sample of women classification is many in this target face image pattern, then target gender classification is women.
If target gender classification is male, then choose in target face image pattern the image pattern of male as similar face image sample
This, if target gender classification is women, then choose in target face image pattern the image pattern of women as similar face figure
Decent.
Illustrate with an object lesson below, how according to the feature of the feature of target facial image and image pattern it
Between cosine similarity comparison identify the sex of object, be described as follows:
For convenience of description, if the feature of 5 image patterns is respectively as follows: sample 1, sample 2, sample 3, sample 4, sample 5,
This target facial image is image 1, and preset number is 3；If the sex classification of sample 1 is male, the sex classification of sample 2 is
Women, the sex classification of sample 3 is male, and the sex classification of sample 4 is women, and the sex classification of sample 5 is male.
By the feature of image 1, feature with sample 1-5 carries out cosine similarity comparison respectively；
The cosine similarity value calculated between image 1 and sample 1 is numerical value 1, calculates the cosine between image 1 and sample 2
Similarity value is numerical value 2, and the cosine similarity value calculated between image 1 and sample 3 is numerical value 3, calculate image 1 and sample 4 it
Between cosine similarity value be numerical value 4, calculating the cosine similarity value between image 1 and sample 5 is numerical value 5, wherein numerical value 5 >
Numerical value 3 > numerical value 4 > numerical value 1 > numerical value 2；
According to the above-mentioned cosine similarity value calculated order from high to low, choose the target facial image sample of preset number
This is sample 5, sample 3 and sample 4；
The sex classification of this target face image pattern is carried out quantity statistics, and the number of samples of male's classification is 2, female
The number of samples of property classification is 1, then male's classification is the second target gender classification；
The target face image pattern determining male's classification is sample 5 and sample 3, and calculates the average of numerical value 5 and numerical value 3
Value 1；
Using meansigma methods 1 and the second target gender classification as this cosine comparison result.
S205, this Sex Discrimination result and this cosine comparison result are carried out Bayes classifier computing, obtain this face
The sex of objects in images.
Bayes classifier is to calculate object by Bayesian formula to belong to the algorithm of which class probability, and Bayes divides
The principle of class device computing is the prior probability by certain object, utilizes Bayesian formula to calculate the posterior probability of this object, i.e.
This object belongs to the probability of a certain class, selects have the classification of maximum a posteriori probability as the classification belonging to this object.
Using this Sex Discrimination result and this cosine comparison result as object in facial image in the embodiment of the present invention
Prior probability, belongs to male's classification and the posterior probability of women classification by what Bayesian formula calculated this object, and chooses
There is the sex classification sex classification as this object of maximum a posteriori probability.
In the embodiment of the present invention, according to each key point of facial image position in this facial image, extract this face
Key position region in image, according to the image of key position regional restructuring face in this facial image, obtains this target face
Image, by prior model in preset degree of depth convolutional neural networks, carries out sex discriminant classification to target facial image, obtains
The Sex Discrimination result that this target facial image is corresponding, by the feature of the feature of this target facial image of comparison Yu image pattern
Between cosine similarity, obtain the cosine comparison result that this target facial image is corresponding, finally by this Sex Discrimination result and
This cosine comparison result carries out Bayes classifier computing, obtains the sex of object in this facial image, so at complicated one-tenth
In the case of Xiang, by key position each in facial image being re-started splicing, and degree of depth convolutional neural networks is differentiated and figure
The aspect ratio of picture is to combining, it is possible to utilize the feature in face each key position region, depth texture, edge to enter with color characteristic
Row assists in identifying, and adds the accuracy of sex identification.
Refer to the structure that Fig. 6, Fig. 6 are the sex identification devices based on facial image that third embodiment of the invention provides
Schematic diagram, for convenience of description, illustrate only the part relevant to the embodiment of the present invention.Fig. 6 example based on facial image
Sex identification device can be the execution master of the gender identification method based on facial image that aforementioned embodiment illustrated in fig. 1 provides
Body, can be a control module in terminal or terminal.The sex identification device based on facial image of Fig. 6 example, mainly
Including: Sex Discrimination module 601, comparison processing module 602 and identification module 603.The most each functional module describes in detail as follows:
Sex Discrimination module 601, for by prior model in preset degree of depth convolutional neural networks, to target face figure
As carrying out sex discriminant classification, obtain the Sex Discrimination result that this target facial image is corresponding；
Comparison processing module 602, between feature and the feature of image pattern by this target facial image of comparison
Cosine similarity, obtain the cosine comparison result that this target facial image is corresponding；
Identification module 603, for this Sex Discrimination result and this cosine comparison result are carried out Bayes classifier computing,
Obtain the sex of object in this facial image.
The image that this target facial image is made up of key position region each in facial image.This each key position region
Can be region and the face contour region of each organ sites of face, herein to target facial image Zhong Ge key position district
Territory does not limits.It should be noted that the relation of facial image and target person image is: this target facial image is by this face
The image of this each key position region composition in image.This facial image is for representing the image of overall face.
Degree of depth convolutional neural networks is a kind of deep neural network with convolutional coding structure, at least include two non-linear can
The convolutional layer of training, two nonlinear fixing convolutional layers and full articulamentum, at least 5 hidden layers, are mainly used in language altogether
Cent analysis and field of image recognition.
Can be obtained by prior model by degree of depth convolutional neural networks training image sample, wherein training image sample with
The process obtaining prior model is the process classified image pattern according to preset classification, and wherein prior model can be used for
Image is carried out the model of discriminant classification.
In field of image recognition, the feature of image is the proper noun of field of image recognition, the extraction of the feature of image
It it is a concept in computer vision and image procossing.
The feature of image pattern can be stored in the memory module of terminal built-in, it is also possible to is stored in cloud server
On.Cosine similarity is as the size weighing two interindividual variations by two vectorial angle cosine values in vector space
Tolerance.It should be noted that object is the people described in this facial image in this facial image.
Bayes classifier is to calculate object by Bayesian formula to belong to the algorithm of which class probability, and Bayes divides
The principle of class device computing is the prior probability by certain object, utilizes Bayesian formula to calculate the posterior probability of this object, i.e.
This object belongs to the probability of a certain class, selects have the classification of maximum a posteriori probability as the classification belonging to this object.The present invention
Embodiment is using this Sex Discrimination result and this cosine comparison result as the prior probability of object in facial image, passes through shellfish
This formula of leaf calculates the male's classification belonging to this object and the posterior probability of women classification, and chooses that to have maximum a posteriori general
The sex classification of rate is as the sex classification of this object.
It should be noted that in the embodiment of the sex identification device based on facial image of figure 6 above example, each merit
Can the division of module be merely illustrative of, can as required in actual application, the configuration requirement of such as corresponding hardware or soft
The convenient consideration of the realization of part, and above-mentioned functions distribution is completed by different functional modules.And, in actual application, this reality
Executing the corresponding functional module in example can be to be realized by corresponding hardware, it is also possible to performed corresponding software by corresponding hardware
Complete.Each embodiment that this specification provides all can apply foregoing description principle, below repeats no more.
In the embodiment of the present invention, Sex Discrimination module 602 is by prior model in preset degree of depth convolutional neural networks, right
Target facial image carries out sex discriminant classification, obtains the Sex Discrimination result that this target facial image is corresponding, wherein said mesh
The image that mark facial image is made up of key position region each in facial image, comparison processing module 603 is by this mesh of comparison
Cosine similarity between feature and the feature of image pattern of mark facial image, obtains the cosine that this target facial image is corresponding
Comparison result, this Sex Discrimination result and this cosine comparison result are carried out Bayes classifier computing, obtain by identification module 604
The sex of object in this facial image, so under complicated imaging contexts, by key position each in facial image again
Splice, and degree of depth convolutional neural networks is differentiated, and the aspect ratio with image is to combining, it is possible to utilize face each key portion
The position feature in region, depth texture, edge assist in identifying with color characteristic, add the accuracy of sex identification.
Refer to Fig. 7, the structural representation of the sex identification device based on facial image that fourth embodiment of the invention provides
Figure, for convenience of description, illustrate only the part relevant to the embodiment of the present invention.The sex based on facial image of Fig. 7 example
Identify that device can be the executive agent of the gender identification method based on facial image that aforementioned embodiment illustrated in fig. 2 provides, can
To be a control module in terminal or terminal.The sex identification device based on facial image of Fig. 7 example, specifically includes that
Extraction module 701, reconstructed module 702, Sex Discrimination module 703, comparison processing module 704, identification module 705, wherein sex
Discrimination module 703 includes: discrimination module 7031 and setting module 7032；Comparison processing module 704 includes: comparing module 7041,
Choose module 7042, statistical module 7043 and module 7044 is set.The most each functional module describes in detail as follows:
Extraction module 701, is used for the position in this facial image of each key point according to facial image, extracts this face
Key position region in image.
Reconstructed module 702, for according to the image of key position regional restructuring face in this facial image, obtains this target
Facial image.
The image that this target facial image is made up of key position region each in facial image.
Alternatively, this device also comprises determining that module is treated for being determined by Face datection and face key point location
Identify the facial image in image；This determines module to be additionally operable to the facial image in this image to be identified to be set as detection zone
Territory.
The image of input is carried out Face datection, then to the image after detection by HAAR grader or DLIB algorithm
Carry out face key point location, wherein being included by the face key point that SDM algorithm positions by SDM algorithm: eyebrow, eyes,
Nose, face and face contour.Certainly Face datection and face key point location can also be realized by other algorithms.
HAAR grader, comprises Adaboost algorithm, and in field of image recognition, grader refers to face and non-face
Carry out the algorithm classified.DLIB is the algorithms library of a kind of C++, can be applicable to Face datection and face key point location.
Fig. 3 is the schematic diagram of Face datection, as it is shown on figure 3, atrous square box is face detection block, and circular expression people
Face, triangular representation animal, Polygons Representation trees, facial image can be extracted in the picture through Face datection, and by people
Face image is arranged in Face datection frame.
Alternatively, this device also include extension module for expanding this detection region according to preset times so that this face
Image includes the image of hair zones.It is hair zones to be extended to that extension module expands this detection region according to preset times
In facial image, the mode expanding detection region can be to be enlarged in whole detection region, it is also possible to is will to detect region
Top, bottom, left and right two parts all expand preset times, it is also possible to only expand the upper and lower in this detection region.This is pre-
Putting multiple and send out the most relevant with hair lengths and brain top, in this enforcement, the numerical value of preferred preset times is 0.15, will
The upper and lower in this detection region respectively expands 0.15 times.Fig. 4 is the schematic diagram of facial image, wherein in Fig. 4 (a) and Fig. 4 (b)
For detection region within rectangle frame, include expanding the facial image before detection region, Fig. 4 in the detection region in Fig. 4 (a)
Include expanding the facial image after detection region in detection region in (b).
Owing to color, length and the hair style of hair are to identify the foundation that sex is critically important, so facial image includes
Hair zones can more accurately identify sex.
Alternatively, this device also include rectification module for eye areas in this detection region is corrected so that should
In facial image, the eyes in this eye areas are on same level line.The mode of correction ocular does not limits, and can pass through
Geometric transformation carrys out correction ocular, it is also possible to changing angle between two eyes and carry out correction ocular, final purpose is by these eyes
Eyes in region are on same level line.
Face key point includes: eyes, nose, face, eyebrow, hair and face contour, and extraction module 701 is by each
Key point position in this facial image, extracts key position region in this facial image.Wherein crucial in this facial image
Area includes: hair zones, brow region, eye areas, nasal area, face region.Reconstructed by reconstructed module 702
The target facial image that each key position region obtains can describe more accurately the hair style of face of different sexes, skin with
And the regional area of the face such as face contour.
As it is shown in figure 5, the schematic diagram that Fig. 5 is the target facial image obtained by each key position regional restructuring.In order to just
In explanation and display, in the schematic diagram shown in Fig. 5 key position region only include eye areas, brow region, nasal area and
Face region, in Figure 5, the schematic diagram shown in Fig. 5 is only a citing to the not display of other key position regions, not
Key position region in the present invention can be constituted and limit.
Sex Discrimination module 703, for by prior model in preset degree of depth convolutional neural networks, to target face figure
As carrying out sex discriminant classification, obtain the Sex Discrimination result that this target facial image is corresponding.
Alternatively, Sex Discrimination module 703 includes: discrimination module 7031 and setting module 7032；
Discrimination module 7031, for by prior model in this degree of depth convolutional neural networks, enters this target facial image
Row discriminant classification, obtains most probable value corresponding to this target facial image and first object corresponding to this most probable value
Other classification；
Setting module 7032, for tying this most probable value and this first object sex classification as described Sex Discrimination
Really.
This first object sex classification is male's classification or women classification.Degree of depth convolutional neural networks is a kind of with convolution
The deep neural network of structure, at least include two non-linear trainable convolutional layers, two nonlinear fixing convolutional layers and
Full articulamentum, at least 5 hidden layers, are mainly used in speech analysis and field of image recognition altogether.Herein to convolutional Neural
Network, the explanation of prior model refer to associated description in first embodiment of the invention, do not repeat.
Comparison processing module 704, between feature and the feature of image pattern by this target facial image of comparison
Cosine similarity, obtain the cosine comparison result that this target facial image is corresponding.
This image pattern is the image pattern being made up of each key position region in overall face image pattern, and wherein this is whole
In body facial image sample, key position region includes: hair zones, brow region, eye areas, nasal area, face district
Territory.The region that in this entirety face image pattern, key position region and key position region in this facial image include is right
Answer, i.e. if this entirety face image pattern includes brow region, then this facial image includes brow region, if this is whole
Body facial image sample includes lens area, then this facial image includes lens area.The image pattern obtained after reconstruct
Composition and the image similarity shown in Fig. 5.The mode in key position region is extracted and from face from overall face image pattern
The extracting mode in image zooming-out key position region is identical, does not repeats.Same, reconstruct obtains the reconstruct of image pattern
The reconstruct mode that mode obtains target facial image with reconstruct is identical, does not repeats.
This cosine similarity is as weighing two interindividual variations by two vectorial angle cosine values in vector space
The tolerance of size.
Alternatively, comparison processing module 704 includes: comparing module 7041, choose module 7042, statistical module 7043, meter
Calculate module 7044 and module 7045 is set.
Comparing module 7041, for by remaining by carrying out between feature and the feature of image pattern of this target facial image
String similarity comparison, obtains multiple cosine similarity values that this target facial image is corresponding；
Choose module 7042, for according to this cosine similarity value order from high to low, choose the target of preset number
Facial image sample；
Statistical module 7043, for carrying out quantity statistics to the sex classification of this target face image pattern, and by quantity
Most sex classifications is as the second target gender classification；
Choosing module 7042, being additionally operable to choose sex classification in this target face image pattern is this second target gender
The similar face image sample of classification, and it is similar to extract the cosine between this similar face image sample to this target facial image
Angle value is as target cosine similarity value；
Computing module 7044, for calculating the meansigma methods of this target cosine similarity value；
Module 7045 is set, for using this second target gender classification and described meansigma methods as this cosine comparison result.
The numerical value of cosine similarity value is the biggest, represents that similarity is the highest.Preset number is arbitrarily chosen, the number certainly chosen
Many, then the last accuracy differentiated will improve.The preferred preset number of the present embodiment is 20.
Statistical module 7043 carries out quantity statistics to the sex classification of this target face image pattern, and by sample size
Many sex classifications are as target gender classification, if the quantity of the sample of male's classification is many in this target face image pattern,
Then target gender classification is male；If in this target face image pattern, the quantity of the sample of women classification is many, then targeted
Other classification is women.If target gender classification is male, then chooses module 7042 and choose male in target face image pattern
Image pattern is as similar face image sample, if target gender classification is women, then chooses module 7042 and chooses target face
In image pattern, the image pattern of women is as similar face image sample.
Identification module 705, for this Sex Discrimination result and this cosine comparison result are carried out Bayes classifier computing,
Obtain the sex of object in this facial image.
Bayes classifier is to calculate object by Bayesian formula to belong to the algorithm of which class probability, and Bayes divides
The principle of class device computing is the prior probability by certain object, utilizes Bayesian formula to calculate the posterior probability of this object, i.e.
This object belongs to the probability of a certain class, selects have the classification of maximum a posteriori probability as the classification belonging to this object.
Using this Sex Discrimination result and this cosine comparison result as object in facial image in the embodiment of the present invention
Prior probability, identification module 705 belongs to male's classification and the posteriority of women classification by what Bayesian formula calculated this object
Probability, and choose the sex classification with maximum a posteriori probability as the sex classification of this object.
In the embodiment of the present invention, extraction module 701 according to each key point of described facial image in described facial image
Position, extract key position region in described facial image, in wherein said facial image, key position region includes: hair
Region, brow region, eye areas, nasal area, face region, reconstructed module 702 is according to key portion in described facial image
The image of position regional restructuring face, obtains described target facial image, Sex Discrimination module 703 is by preset degree of depth convolution
Prior model in neutral net, carries out sex discriminant classification to target facial image, obtains described target facial image corresponding
Sex Discrimination result, the image that wherein said target facial image is made up of key position region each in facial image, then
Comparison processing module 704 is similar by the cosine between feature and the feature of image pattern of target facial image described in comparison
Degree, obtains the cosine comparison result that described target facial image is corresponding, and last identification module 705 is by this Sex Discrimination result and is somebody's turn to do
Cosine comparison result carries out Bayes classifier computing, obtains the sex of object in this facial image, so in complicated imaging
In the case of, by key position each in facial image being re-started splicing, and degree of depth convolutional neural networks is differentiated and image
Aspect ratio to combining, it is possible to utilize the feature in face each key position region, depth texture, edge to carry out with color characteristic
Assist in identifying, add the accuracy of sex identification.
Fig. 8 is the hardware of the electronic equipment of the gender identification method based on facial image that fifth embodiment of the invention provides
Structural representation, as shown in Figure 8, this electronic equipment includes:
One or more processors 810 and memorizer 820, in Fig. 8 as a example by a processor 810.
This electronic equipment can also include: input equipment 830 and output device 840.
Processor 810, memorizer 820, input equipment 830 and output device 840 can be by bus or other modes
Connect, in Fig. 8 as a example by being connected by bus 850.
Memorizer 820, as a kind of non-volatile computer readable storage medium storing program for executing, can be used for storing non-volatile software journey
Sequence, non-volatile computer executable program and module, program as corresponding in the date storage method in the embodiment of the present application
Instruction/module (such as, the Sex Discrimination module 601 shown in accompanying drawing 6, comparison processing module 602 and identification module 603).Process
Non-volatile software program, instruction and the module that device 810 is stored in memorizer 820 by operation, thus perform server
Various functions application and data process, i.e. realize the gender identification method based on facial image of said method embodiment.
Memorizer 820 can include storing program area and storage data field, and wherein, storage program area can store operation system
Application program required for system, at least one function；Storage data field can store and fill according to sex identification based on facial image
The data etc. that the use put is created.Additionally, memorizer 820 can include high-speed random access memory, it is also possible to include non-
Volatile memory, for example, at least one disk memory, flush memory device or other non-volatile solid state memory parts.?
In some embodiments, memorizer 820 is optional includes the memorizer remotely located relative to processor 810, these remote memories
Sex identification device based on facial image can be connected to by network.The example of above-mentioned network includes but not limited to interconnection
Net, intranet, LAN, mobile radio communication and combinations thereof.
Input equipment 830 can receive numeral or the character information of input, and produces and sex based on facial image knowledge
The user setup of other device and function control relevant key signals input.Output device 840 can include that the displays such as display screen set
Standby.
One or more module stores is in described memorizer 820, when by one or more processor
During 810 execution, perform the gender identification method based on facial image in above-mentioned any means embodiment.
The said goods can perform the method that the embodiment of the present application is provided, and possesses the corresponding functional module of execution method and has
Benefit effect.The ins and outs of the most detailed description, can be found in the method that the embodiment of the present application is provided.
The electronic equipment of the embodiment of the present invention exists in a variety of forms, includes but not limited to:
(1) mobile communication equipment: the feature of this kind equipment is to possess mobile communication function, and to provide speech, data
Communication is main target.This Terminal Type includes: smart mobile phone, multimedia handset, functional mobile phone, and low-end mobile phone etc..
(2) super mobile personal computer equipment: this kind equipment belongs to the category of personal computer, has calculating and processes merit
Can, the most also possess mobile Internet access characteristic.This Terminal Type includes: palm PC (PDA Personal Digital
Assistant), mobile internet device (MID, Mobile Internet Device) and Ultra-Mobile PC
(UMPC, Ultra-mobile Personal Computer) equipment etc..
(3) portable entertainment device: this kind equipment can show and play content of multimedia.This kind equipment includes: audio frequency,
Video player, handheld device, e-book, and intelligent toy and portable car-mounted navigator.
(4) server: providing the equipment of the service of calculating, the composition of server includes that processor, hard disk, internal memory, system are total
Lines etc., server is similar with general computer architecture, but owing to needing to provide highly reliable service, is therefore processing energy
The aspects such as power, stability, reliability, safety, extensibility, manageability require higher.
(5) other have the electronic installation of data interaction function.
In multiple embodiments provided herein, it should be understood that disclosed system, terminal and method, permissible
Realize by another way.Such as, terminal embodiment described above is only schematically, such as, and described module
Dividing, be only a kind of logic function and divide, actual can have other dividing mode, the most multiple modules or assembly when realizing
Can in conjunction with or be desirably integrated into another system, or some features can be ignored, or does not performs.Another point, shown or
The coupling each other discussed or direct-coupling or communication linkage can be by some interfaces, the INDIRECT COUPLING or logical of module
Letter link, can be electrical, machinery or other form.
The described module illustrated as separating component can be or may not be physically separate, shows as module
The parts shown can be or may not be physical module, i.e. may be located at a place, or can also be distributed to multiple
On mixed-media network modules mixed-media.Some or all of module therein can be selected according to the actual needs to realize the mesh of the present embodiment scheme
's.
It addition, each functional module in each embodiment of the present invention can be integrated in a processing module, it is also possible to
It is that modules is individually physically present, it is also possible to two or more modules are integrated in a module.Above-mentioned integrated mould
Block both can realize to use the form of hardware, it would however also be possible to employ the form of software function module realizes.
If described integrated module realizes and as independent production marketing or use using the form of software function module
Time, can be stored in a computer read/write memory medium.Based on such understanding, technical scheme is substantially
The part that in other words prior art contributed or this technical scheme completely or partially can be with the form of software product
Embodying, this computer software product is stored in a storage medium, including some instructions with so that a computer
Equipment (can be personal computer, server, or the network equipment etc.) performs the complete of method described in each embodiment of the present invention
Portion or part steps.And aforesaid storage medium includes: USB flash disk, portable hard drive, read only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
It should be noted that for aforesaid each method embodiment, in order to simplicity describes, therefore it is all expressed as a series of
Combination of actions, but those skilled in the art should know, the present invention is not limited by described sequence of movement because
According to the present invention, some step can use other order or carry out simultaneously.Secondly, those skilled in the art also should know
Knowing, it might not be all this that embodiment described in this description belongs to preferred embodiment, involved action and module
Bright necessary.
In the above-described embodiments, the description to each embodiment all emphasizes particularly on different fields, and does not has the portion described in detail in certain embodiment
Point, may refer to the associated description of other embodiments.
It is more than to gender identification method based on facial image provided by the present invention and the description of device, for ability
The technical staff in territory, according to the thought of the embodiment of the present invention, the most all will change,
To sum up, this specification content should not be construed as limitation of the present invention.
Claims (10)
1. a gender identification method based on facial image, it is characterised in that including:
By prior model in preset degree of depth convolutional neural networks, target facial image is carried out sex discriminant classification, obtains
The Sex Discrimination result that described target facial image is corresponding, wherein said target facial image is by key portion each in facial image
The image of region, position composition；
By the cosine similarity between feature and the feature of image pattern of target facial image described in comparison, obtain described mesh
The cosine comparison result that mark facial image is corresponding；
Described Sex Discrimination result and described cosine comparison result are carried out Bayes classifier computing, obtains described facial image
The sex of middle object.
Method the most according to claim 1, it is characterised in that in described facial image, each key position region includes: head
Send out region, brow region, eye areas, nasal area, face region.
Method the most according to claim 2, it is characterised in that described by priori in preset degree of depth convolutional neural networks
Model, carries out sex discriminant classification to target facial image, obtains the Sex Discrimination result that described target facial image is corresponding, tool
Body includes:
By prior model in described degree of depth convolutional neural networks, described target facial image is carried out discriminant classification, obtains institute
State most probable value corresponding to target facial image and first object sex classification corresponding to described most probable value；
Using described most probable value and described first object sex classification as described Sex Discrimination result.
Method the most according to claim 2, it is characterised in that described image pattern is by each in overall face image pattern
The image pattern of key position region composition, in wherein said overall face image pattern, key position region includes: hair district
Territory, brow region, eye areas, nasal area, face region.
Method the most according to claim 4, it is characterised in that described by the feature of target facial image described in comparison with
Cosine similarity between the feature of image pattern, obtains the cosine comparison result that described target facial image is corresponding, specifically wraps
Include:
By by carrying out cosine similarity comparison between feature and the feature of image pattern of described target facial image, obtaining many
Individual cosine similarity value；
According to described cosine similarity value order from high to low, choose the target face image pattern of preset number；
The sex classification of described target face image pattern is carried out quantity statistics, and using sex classifications most for quantity as
Two target gender classifications；
The similar face image that sex classification is described second target gender classification is chosen in described target face image pattern
Sample, and extract cosine similarity value between described similar face image sample and described target facial image as more than target
String Similarity value；
Calculate the meansigma methods of described target cosine similarity value；
Using described second target gender classification and described meansigma methods as described cosine comparison result.
6. a sex identification device based on facial image, it is characterised in that including:
Sex Discrimination module, for by prior model in preset degree of depth convolutional neural networks, is carried out target facial image
Gender Classification differentiate, obtain the Sex Discrimination result that described target facial image is corresponding, wherein said target facial image be by
The image of each key position region composition in facial image；
Comparison processing module, for by the cosine between feature and the feature of image pattern of target facial image described in comparison
Similarity, obtains the cosine comparison result that described target facial image is corresponding；
Identification module, for described Sex Discrimination result and described cosine comparison result are carried out Bayes classifier computing,
The sex of object in described facial image.
Device the most according to claim 6, it is characterised in that in described facial image, each key position region includes: head
Send out region, brow region, eye areas, nasal area, face region.
Device the most according to claim 7, it is characterised in that described Sex Discrimination module includes:
Discrimination module, for by prior model in described degree of depth convolutional neural networks, is carried out described target facial image point
Class differentiates, obtains most probable value corresponding to described target facial image and first object corresponding to described most probable value
Other classification；
Setting module, for using described most probable value and described first object sex classification as described Sex Discrimination result.
Device the most according to claim 7, it is characterised in that described image pattern is by each in overall face image pattern
The image pattern of key position region composition, in wherein said overall face image pattern, key position region includes: hair district
Territory, brow region, eye areas, nasal area, face region.
Device the most according to claim 8, it is characterised in that described comparison processing module includes:
Comparing module, for by similar by carrying out cosine between feature and the feature of image pattern of described target facial image
Degree comparison, obtains multiple cosine similarity value；
Choose module, for according to described cosine similarity value order from high to low, choosing the target face figure of preset number
Decent；
Statistical module is for carrying out quantity statistics to the sex classification of described target face image pattern and quantity is most
Sex classification is as the second target gender classification；
Described choosing module, being additionally operable to choose sex classification in described target face image pattern is described second target gender
The similar face image sample of classification, and extract the cosine between described similar face image sample and described target facial image
Similarity value is as target cosine similarity value；
Computing module, for calculating the meansigma methods of described target cosine similarity value；
Module is set, for using described second target gender classification and described meansigma methods as described cosine comparison result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610698494.4A CN106326857A (en) | 2016-08-19 | 2016-08-19 | Gender identification method and gender identification device based on face image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610698494.4A CN106326857A (en) | 2016-08-19 | 2016-08-19 | Gender identification method and gender identification device based on face image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106326857A true CN106326857A (en) | 2017-01-11 |
Family
ID=57741455
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610698494.4A Pending CN106326857A (en) | 2016-08-19 | 2016-08-19 | Gender identification method and gender identification device based on face image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106326857A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107578029A (en) * | 2017-09-21 | 2018-01-12 | 北京邮电大学 | Method, apparatus, electronic equipment and the storage medium of area of computer aided picture certification |
| CN107886474A (en) * | 2017-11-22 | 2018-04-06 | 北京达佳互联信息技术有限公司 | Image processing method, device and server |
| CN107958245A (en) * | 2018-01-12 | 2018-04-24 | 上海正鹏信息科技有限公司 | A kind of gender classification method and device based on face characteristic |
| CN108122001A (en) * | 2017-12-13 | 2018-06-05 | 北京小米移动软件有限公司 | Image-recognizing method and device |
| CN108154169A (en) * | 2017-12-11 | 2018-06-12 | 北京小米移动软件有限公司 | Image processing method and device |
| CN108664948A (en) * | 2018-05-21 | 2018-10-16 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN109271957A (en) * | 2018-09-30 | 2019-01-25 | 厦门市巨龙信息科技有限公司 | Face gender identification method and device |
| CN109493490A (en) * | 2018-11-27 | 2019-03-19 | 电卫士智能电器（北京）有限公司 | Electricity consumption user right judgment method and device |
| CN109584417A (en) * | 2018-11-27 | 2019-04-05 | 电卫士智能电器（北京）有限公司 | Door-access control method and device |
| CN110210293A (en) * | 2018-10-30 | 2019-09-06 | 上海市服装研究所有限公司 | A kind of gender identification method based on three-dimensional data and face-image |
| CN110751215A (en) * | 2019-10-21 | 2020-02-04 | 腾讯科技（深圳）有限公司 | Image identification method, device, equipment, system and medium |
| CN110826525A (en) * | 2019-11-18 | 2020-02-21 | 天津高创安邦技术有限公司 | Face recognition method and system |
| CN110880011A (en) * | 2018-09-05 | 2020-03-13 | 宏达国际电子股份有限公司 | Image segmentation method, device and non-transitory computer readable medium thereof |
| CN111353943A (en) * | 2018-12-20 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Face image recovery method and device and readable storage medium |
| CN111951267A (en) * | 2020-09-08 | 2020-11-17 | 南方科技大学 | Gender judgment method, device, server and storage medium based on neural network |
| CN112882042A (en) * | 2021-01-14 | 2021-06-01 | 天津市水产研究所 | Marine ranching seabed telemetering and identifying method based on acoustic data |
-
2016
- 2016-08-19 CN CN201610698494.4A patent/CN106326857A/en active Pending
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107578029A (en) * | 2017-09-21 | 2018-01-12 | 北京邮电大学 | Method, apparatus, electronic equipment and the storage medium of area of computer aided picture certification |
| CN107578029B (en) * | 2017-09-21 | 2020-03-27 | 北京邮电大学 | Computer-aided picture authentication method and device, electronic equipment and storage medium |
| CN107886474A (en) * | 2017-11-22 | 2018-04-06 | 北京达佳互联信息技术有限公司 | Image processing method, device and server |
| CN108154169A (en) * | 2017-12-11 | 2018-06-12 | 北京小米移动软件有限公司 | Image processing method and device |
| CN108122001A (en) * | 2017-12-13 | 2018-06-05 | 北京小米移动软件有限公司 | Image-recognizing method and device |
| CN108122001B (en) * | 2017-12-13 | 2022-03-11 | 北京小米移动软件有限公司 | Image recognition method and device |
| CN107958245A (en) * | 2018-01-12 | 2018-04-24 | 上海正鹏信息科技有限公司 | A kind of gender classification method and device based on face characteristic |
| CN108664948A (en) * | 2018-05-21 | 2018-10-16 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN110880011A (en) * | 2018-09-05 | 2020-03-13 | 宏达国际电子股份有限公司 | Image segmentation method, device and non-transitory computer readable medium thereof |
| CN109271957B (en) * | 2018-09-30 | 2020-10-20 | 厦门市巨龙信息科技有限公司 | Face gender identification method and device |
| CN109271957A (en) * | 2018-09-30 | 2019-01-25 | 厦门市巨龙信息科技有限公司 | Face gender identification method and device |
| CN110210293A (en) * | 2018-10-30 | 2019-09-06 | 上海市服装研究所有限公司 | A kind of gender identification method based on three-dimensional data and face-image |
| CN109584417A (en) * | 2018-11-27 | 2019-04-05 | 电卫士智能电器（北京）有限公司 | Door-access control method and device |
| CN109493490A (en) * | 2018-11-27 | 2019-03-19 | 电卫士智能电器（北京）有限公司 | Electricity consumption user right judgment method and device |
| CN111353943A (en) * | 2018-12-20 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Face image recovery method and device and readable storage medium |
| CN111353943B (en) * | 2018-12-20 | 2023-12-26 | 杭州海康威视数字技术股份有限公司 | Face image recovery method and device and readable storage medium |
| CN110751215A (en) * | 2019-10-21 | 2020-02-04 | 腾讯科技（深圳）有限公司 | Image identification method, device, equipment, system and medium |
| CN110826525A (en) * | 2019-11-18 | 2020-02-21 | 天津高创安邦技术有限公司 | Face recognition method and system |
| CN110826525B (en) * | 2019-11-18 | 2023-05-26 | 天津高创安邦技术有限公司 | Face recognition method and system |
| CN111951267A (en) * | 2020-09-08 | 2020-11-17 | 南方科技大学 | Gender judgment method, device, server and storage medium based on neural network |
| CN112882042A (en) * | 2021-01-14 | 2021-06-01 | 天津市水产研究所 | Marine ranching seabed telemetering and identifying method based on acoustic data |
| CN112882042B (en) * | 2021-01-14 | 2022-06-21 | 天津市水产研究所 | Marine ranching seabed telemetering and identifying method based on acoustic data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106326857A (en) | Gender identification method and gender identification device based on face image | |
| CN106295591A (en) | Gender identification method based on facial image and device | |
| CN106469298A (en) | Age recognition methodss based on facial image and device | |
| US12039454B2 (en) | Microexpression-based image recognition method and apparatus, and related device | |
| US10867405B2 (en) | Object learning and recognition method and system | |
| CN110555481B (en) | Portrait style recognition method, device and computer readable storage medium | |
| US12087097B2 (en) | Image recognition method and apparatus, computer-readable storage medium, and electronic device | |
| US9965705B2 (en) | Systems and methods for attention-based configurable convolutional neural networks (ABC-CNN) for visual question answering | |
| CN106068514B (en) | System and method for identifying face in free media | |
| CN109815826B (en) | Method and device for generating face attribute model | |
| Li et al. | Universal sketch perceptual grouping | |
| CN108388876A (en) | A kind of image-recognizing method, device and relevant device | |
| CN106407911A (en) | Image-based eyeglass recognition method and device | |
| US20140143183A1 (en) | Hierarchical model for human activity recognition | |
| CN108229268A (en) | Expression Recognition and convolutional neural networks model training method, device and electronic equipment | |
| CN109145871B (en) | Psychological behavior recognition method, device and storage medium | |
| CN108182409A (en) | Biopsy method, device, equipment and storage medium | |
| CN106909938B (en) | Visual angle independence behavior identification method based on deep learning network | |
| US20230095182A1 (en) | Method and apparatus for extracting biological features, device, medium, and program product | |
| CN107609563A (en) | Picture semantic describes method and device | |
| CN110287848A (en) | The generation method and device of video | |
| CN113205017A (en) | Cross-age face recognition method and device | |
| EP4435660A1 (en) | Target detection method and apparatus | |
| Zhang et al. | Classroom student posture recognition based on an improved high-resolution network | |
| Ji | Combining knowledge with data for efficient and generalizable visual learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170111 |
|
| WD01 | Invention patent application deemed withdrawn after publication |