CN104899579A - Face recognition method and face recognition device - Google Patents
Face recognition method and face recognition device Download PDFInfo
- Publication number
- CN104899579A CN104899579A CN201510369876.8A CN201510369876A CN104899579A CN 104899579 A CN104899579 A CN 104899579A CN 201510369876 A CN201510369876 A CN 201510369876A CN 104899579 A CN104899579 A CN 104899579A
- Authority
- CN
- China
- Prior art keywords
- facial image
- identified
- learning model
- image sample
- degree
- 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
- G06V40/168—Feature extraction; Face representation
-
- 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
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Abstract
The invention provides a face recognition method. The face recognition method includes: extracting a to-be-recognized face image and a face feature vector of a face image sample according to a preset depth learning model; computing a similarity value of the to-be-recognized face image and the face image sample based on the extracted face feature vector; subjecting the to-be-recognized face image to face recognition according to the computed similarity value. The face recognition method has the advantage that face features can be extracted from the face image by the aid of the depth learning model, so that face recognition accuracy can be improved.
Description
Technical field
The disclosure relates to communication field, particularly relates to face identification method and device.
Background technology
Recognition of face, has normally come based on the face characteristic extracted from facial image; Such as, in traditional realization, from facial image, manually can extract face characteristic, and after dimensionality reduction being carried out to the face characteristic extracted in conjunction with specific algorithm, carry out similarity measurement and finally obtain face recognition result.Visible, carrying out in face recognition process, the face characteristic extracted is depended in the accuracy of identification, therefore how to be optimized the process extracting face characteristic, to improve the accuracy of recognition of face, become the emphasis that field of face identification is paid close attention at present.
Summary of the invention
For overcoming Problems existing in correlation technique, the disclosure provides a kind of face identification method and device.
According to the first aspect of disclosure embodiment, provide a kind of face identification method, described method comprises:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
Optionally, before the degree of deep learning model that described basis is preset extracts the face feature vector of facial image to be identified and facial image sample, described method also comprises:
Facial image sample based on predetermined number is trained described degree of deep learning model, to determine the optimal weight parameter of each node in described degree of deep learning model.
Optionally, before the described face feature vector based on extracting calculates the similarity value of described facial image to be identified and described facial image sample, described method comprises:
According to preset algorithm, dimension-reduction treatment is carried out to the described face feature vector extracted.
Optionally, described degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net;
The face feature vector that the degree of deep learning model that described basis is preset extracts facial image to be identified and facial image sample comprises:
Described facial image to be identified and described facial image sample are carried out features training successively in multiple basic units that input picture comprises respectively in described degree of deep learning model;
After training completes, extract full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample.
Optionally, the similarity value that the described face feature vector based on extracting calculates described facial image to be identified and described facial image sample comprises:
Vector distance between the face feature vector calculating described facial image to be identified and described facial image sample; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
According to the similarity switching strategy preset, the described vector distance calculated is converted to corresponding similarity value.
Optionally, the described similarity value that described basis calculates is carried out recognition of face to described facial image to be identified and is comprised:
Judge whether the described similarity value calculated reaches threshold value;
When described similarity value reaches threshold value, confirm that described facial image to be identified is identical with described facial image sample, and described facial image sample is exported as recognition result.
Optionally, described vector distance comprises COS distance or Euclidean distance.
Optionally, described facial image to be identified and facial image sample standard deviation have carried out dimension normalization process in advance centered by eyes.
According to the second aspect of disclosure embodiment, provide a kind of face identification device, described device comprises:
Extraction module, for extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset;
Computing module, for calculating the similarity value of described facial image to be identified and described facial image sample based on the face feature vector extracted;
Identification module, for carrying out recognition of face according to the described similarity value calculated to described facial image to be identified.
Optionally, described device also comprises:
Training module, for before extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, facial image sample based on predetermined number is trained described degree of deep learning model, to determine the optimal weight parameter of each node in described degree of deep learning model.
Optionally, described device also comprises:
Dimensionality reduction module, for before the similarity value calculating described facial image to be identified and described facial image sample based on the face feature vector extracted, carries out dimension-reduction treatment according to preset algorithm to the described face feature vector extracted.
Optionally, described degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net;
Described extraction module comprises:
Training submodule, for carrying out features training using described facial image to be identified and described facial image sample successively in multiple basic units that input picture comprises respectively in described degree of deep learning model;
Extract submodule, for after training completes, extract full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample.
Optionally, described computing module comprises:
Calculating sub module, for calculate described facial image to be identified and described facial image sample face feature vector between vector distance; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
Transform subblock, for being converted to corresponding similarity value according to the similarity switching strategy preset by the described vector distance calculated.
Optionally, described identification module comprises:
Judge submodule, for judging whether the described similarity value calculated reaches threshold value;
Output sub-module, for when described similarity value reaches threshold value, confirms that described facial image to be identified is identical with described facial image sample, and is exported as recognition result by described facial image sample.
Optionally, described vector distance comprises COS distance or Euclidean distance.
Optionally, described facial image to be identified and facial image sample standard deviation have carried out dimension normalization process in advance centered by eyes.
According to the third aspect of disclosure embodiment, a kind of face identification device is provided, comprises:
Processor;
For the storer of storage of processor executable instruction;
Wherein, described processor is configured to:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
In above embodiment of the present disclosure, by extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, and the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted, then according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified, due in the process of recognition of face, degree of deep learning model is utilized automatically from facial image, to extract face characteristic, instead of traditional artificial mode extracting face characteristic from facial image, therefore the degree of accuracy of recognition of face can be improved.
Should be understood that, it is only exemplary and explanatory that above general description and details hereinafter describe, and can not limit the disclosure.
Accompanying drawing explanation
Accompanying drawing to be herein merged in instructions and to form the part of this instructions, shows and meets embodiment of the present disclosure, and is used from instructions one and explains principle of the present disclosure.
Fig. 1 is the schematic flow sheet of a kind of face identification method according to an exemplary embodiment;
Fig. 2 is the schematic flow sheet of the another kind of face identification method according to an exemplary embodiment;
Fig. 3 is the schematic diagram of a kind of depth recognition model according to an exemplary embodiment;
Fig. 4 is the schematic block diagram of a kind of face identification device according to an exemplary embodiment;
Fig. 5 is the schematic block diagram of the another kind of face identification device according to an exemplary embodiment;
Fig. 6 is the schematic block diagram of the another kind of face identification device according to an exemplary embodiment;
Fig. 7 is the schematic block diagram of the another kind of face identification device according to an exemplary embodiment;
Fig. 8 is the schematic block diagram of the another kind of face identification device according to an exemplary embodiment;
Fig. 9 is the schematic block diagram of the another kind of face identification device according to an exemplary embodiment;
Figure 10 is a kind of structural representation for described face identification device according to an exemplary embodiment.
Embodiment
Here will be described exemplary embodiment in detail, its sample table shows in the accompanying drawings.When description below relates to accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawing represents same or analogous key element.Embodiment described in following exemplary embodiment does not represent all embodiments consistent with the disclosure.On the contrary, they only with as in appended claims describe in detail, the example of apparatus and method that aspects more of the present disclosure are consistent.
The term used in the disclosure is only for the object describing specific embodiment, and the not intended to be limiting disclosure." one ", " described " and " being somebody's turn to do " of the singulative used in disclosure and the accompanying claims book is also intended to comprise most form, unless context clearly represents other implications.It is also understood that term "and/or" used herein refer to and comprise one or more project of listing be associated any or all may combine.
Term first, second, third, etc. may be adopted although should be appreciated that to describe various information in the disclosure, these information should not be limited to these terms.These terms are only used for the information of same type to be distinguished from each other out.Such as, when not departing from disclosure scope, the first information also can be called as the second information, and similarly, the second information also can be called as the first information.Depend on linguistic context, word as used in this " if " can be construed as into " ... time " or " when ... time " or " in response to determining ".
In the related, when carrying out recognition of face, usually from facial image, manually can extract face characteristic, and after dimensionality reduction being carried out to the face characteristic extracted in conjunction with specific algorithm, carry out similarity measurement and finally obtain face recognition result.
Such as, in tradition realizes, the textural characteristics such as gabor feature or LBP feature can be extracted from facial image, again in conjunction with PCA (Principal Components Analysis, principal component analysis (PCA)) algorithm and LDA (Linear Discriminant Analysis, linear discriminant analysis) algorithm carries out projection matrix training to the textural characteristics extracted, dimensionality reduction is carried out to the textural characteristics extracted, then similarity measurement is carried out to the textural characteristics after dimensionality reduction and obtain face recognition result.
But, carrying out in face recognition process, because the face characteristic extracted is depended in the accuracy identified, therefore by manually extracting face characteristic from facial image, the problem of the accuracy deficiency of recognition of face may caused.
In view of this, the disclosure proposes a kind of face identification method, by extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, and the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted, then according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified, due in the process of recognition of face, degree of deep learning model is utilized automatically from facial image, to extract face characteristic, instead of traditional artificial mode extracting face characteristic from facial image, therefore the degree of accuracy of recognition of face can be improved.
As shown in Figure 1, Fig. 1 is a kind of face identification method according to an exemplary embodiment, and this face identification method is used for service end, comprises the following steps:
In a step 101, the face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
In a step 102, the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
In step 103, according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified.
In the present embodiment, service end can be server, server cluster or the cloud platform that user oriented provides recognition of face to serve.The degree of deep learning model preset can be the degree of deep learning model based on multilayer neural network.Based in the degree of deep learning model of multilayer neural network, usually defeatedly can comprise multiple basic unit, each basic unit can be used as the independently local feature of feature extraction layer to facial image and extracts.
Such as, when realizing, described multilayer neural network can adopt convolutional neural networks comparatively ripe at present, based in the degree of deep learning model of convolutional neural networks, can comprise input layer, multiple convolutional layer, full articulamentum and output layer for carrying out feature extraction.Input layer is used for providing input channel for face image pattern; Convolutional layer can carry out training as independently feature extraction layer to the local feature of facial image and extract, full articulamentum can to each convolutional layer train the local feature extracted to integrate, being trained by each convolutional layer the characteristics of image that extracts to connect is an one-dimensional vector; Output layer is for exporting the classification results of the face figure sample to input.
Be be described based on the degree of deep learning model of convolutional neural networks for described degree of deep learning model below.
Based in the degree of deep learning model of convolutional neural networks, each basic unit comprises the node of some linear one dimensional arrangement, between basic unit and the node of basic unit is wherein to be in a kind of full state be connected, and the connection between node has a weight parameter usually.In an initial condition, the weight parameter connected between node is default value, therefore before formal this degree of deep learning model of use carries out face feature vector extraction to facial image to be identified and facial image sample, need to train this degree of deep learning model, to determine the optimal weight parameter of the connection between each node.
Wherein, when training this degree of deep learning model, the facial image sample of predetermined number can be prepared, and by user, these facial image samples be classified; Such as, 50,000 facial image samples can be prepared, then the user belonged to according to these faces figure sample classifies to these 50,000 facial image samples, and each facial image sample of classifying is demarcated according to affiliated user, such as, each classification sorted can be demarcated as Zhang San, Li Si, king five etc. respectively, each user possesses 10 ~ 1000 pictures do not waited, and the facial image sample standard deviation now in each classification belongs to same user.
After the facial image sample classification of the predetermined number prepared completes, now can using this degree of deep learning model as disaggregated model, these facial image samples are input in this degree of deep learning model as training sample and train, and according to the classification results that degree of deep learning model exports, constantly the weight parameter of the connection in this each basic unit of degree of deep learning model between node is adjusted.In continuous adjustment process, this degree of deep learning model is after the training sample based on input is trained, and compared with the classification results that classification results and the user of output demarcate, accuracy will improve gradually.Meanwhile, user can pre-set an accuracy threshold value, in continuous adjustment process, if compared with the classification results that the classification results that this degree of deep learning model exports and user demarcate, after accuracy reaches the accuracy threshold value pre-set, the weight parameter now connected between each base level nodes in this degree of deep learning model is optimal weight parameter, can think that this degree of deep learning model has been trained complete.
For the degree of deep learning model that training is complete, directly can use this degree of deep learning model, facial image to be identified and facial image sample be carried out to the extraction of face feature vector.
When realizing, service end can create a face image sample data storehouse in advance in this locality, each facial image sample standard deviation in this database can when carrying out recognition of face to facial image to be identified, as the object of reference of comparing with this facial image to be identified.When carrying out face characteristic for the facial image sample in database and extracting, facial image sample in database can be carried out features training successively by service end in multiple convolutional layers that input picture comprises in this degree of deep learning model, after each convolutional layer has all been trained, the face feature vector of proper vector as this facial image sample of full articulamentum output can be extracted.
Due in this degree of deep learning model, full articulamentum can to each convolutional layer train the local feature extracted to integrate, therefore the proper vector exported by full articulamentum is as the face feature vector of this facial image sample, the global characteristics of this face figure sample after being integrated by full articulamentum can be obtained, when using global characteristics to carry out measuring similarity, face recognition result will be more accurate.
Certainly, when realizing, proper vector that any one in multiple convolutional layers of this degree of deep learning model convolutional layer of specifying the exports face feature vector as this facial image sample also can be extracted.By the proper vector that any one convolutional layer of specifying in multiple convolutional layer exports, the local feature of this face figure sample can be obtained.
In the present embodiment, when carrying out face characteristic for facial image to be identified and extracting, service end still can according to identical processing mode, this facial image to be identified is carried out features training successively in multiple convolutional layers that input picture comprises in this degree of deep learning model, after each convolutional layer has all been trained, the face feature vector of proper vector as this facial image to be identified of the convolutional layer output of specifying in full articulamentum or multiple convolutional layer can be extracted.
What deserves to be explained is, service end using the facial image sample in database and facial image to be identified as before input picture carries out features training in this degree of deep learning model, dimension normalization process can be carried out to the facial image sample in database and facial image to be identified centered by eyes, be processed into the image of unified size, compare in the recognition of face stage to facilitate.Wherein, the image size adopted during normalized can be arranged according to the demand of reality; Such as, when being normalized, the image being arranged to 224*224 can be unified.
When after service end to extract facial image sample in database and facial image to be identified face feature vector by this degree of deep learning model, measuring similarity can be carried out based on the face feature vector extracted, calculate face recognition result.
In the present embodiment, when carrying out measuring similarity based on the face feature vector extracted, the vector distance of the face feature vector of the facial image sample in facial image to be identified and database can be utilized to characterize the similarity of the two.
When realizing, service end can calculate the vector distance between the face feature vector of facial image sample in the face feature vector of facial image to be identified and database successively, then the vector distance calculated is converted to corresponding similarity value according to the similarity switching strategy preset.Such as, this similarity switching strategy can be that service end sets up the corresponding relation list of a vector distance and similarity value in advance according to the relation between proper vector and similarity, multiple different similarity grade can be divided into according to the vector distance threshold value preset in this corresponding relation list, and a corresponding similarity value is set for each similarity grade, because the similarity of the vector distance between proper vector usually and between proper vector is inversely proportional to, therefore when vector distance more hour, similarity value is higher, when vector distance is larger, similarity value is lower.Directly just can obtain the similarity value corresponding with the vector distance calculated by this corresponding relation list of inquiry in this way.Wherein, this vector distance can be COS distance, can be also Euclidean distance, not be particularly limited in the present embodiment.
Wherein, service end, before the vector distance of face feature vector calculating described facial image to be identified and described facial image sample, can also carry out dimension-reduction treatment according to the algorithm preset to the face feature vector of the two.Such as, projection matrix training can be carried out by PCA algorithm and LDA algorithm to the face feature vector of the two, dimensionality reduction is carried out to the face feature vector of the two.In this way, the calculated amount of service end can not only be reduced, the redundant information in face feature vector can also be removed, retain the useful information in face feature vector.
After service end similarity value corresponding to the vector distance calculated converts to, can judge whether this similarity value reaches similarity threshold, if when this similarity value reaches similarity threshold, now service end can confirm that this facial image to be identified is identical facial image with this facial image sample, and is exported as recognition result by this facial image sample.Certainly, if when this similarity value does not reach similarity threshold, now service end can confirm that this facial image to be identified is not identical facial image with this facial image sample, now service end can repeat above process, continue the similarity value calculating next face image pattern in this facial image to be identified and database, until find identical facial image, or travel through when whole database does not find facial image to be identified with this identical facial image and stop.
In the embodiment above, by extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, and the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted, then according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified, due in the process of recognition of face, degree of deep learning model is utilized automatically from facial image, to extract face characteristic, instead of traditional artificial mode extracting face characteristic from facial image, therefore the degree of accuracy of recognition of face can be improved.
As shown in Figure 2, Fig. 2 is a kind of face identification method according to an exemplary embodiment, is applied in service end, comprises the following steps:
In step 201, described facial image to be identified and described facial image sample are carried out features training successively in multiple basic units that input picture comprises respectively in described degree of deep learning model; Wherein said degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net;
In step 202., after training completes, extract full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample;
In step 203, the vector distance between the face feature vector calculating described facial image to be identified and described facial image sample; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
In step 204, according to the similarity switching strategy preset, the described vector distance calculated is converted to corresponding similarity value;
In step 205, judge whether the described similarity value calculated reaches threshold value;
In step 205, when described similarity value reaches threshold value, confirm that described facial image to be identified is identical with described facial image sample, and described facial image sample is exported as recognition result.
In the present embodiment, service end can be server, server cluster or the cloud platform that user oriented provides recognition of face to serve.The degree of deep learning model preset can be the degree of deep learning model based on multilayer neural network.Based in the degree of deep learning model of multilayer neural network, usually defeatedly can comprise multiple basic unit, each basic unit can be used as the independently local feature of feature extraction layer to facial image and extracts.
Such as, when realizing, described multilayer neural network can adopt convolutional neural networks comparatively ripe at present, based in the degree of deep learning model of convolutional neural networks, can comprise input layer, multiple convolutional layer, full articulamentum and output layer for carrying out feature extraction.Input layer is used for providing input channel for face image pattern; Convolutional layer can carry out training as independently feature extraction layer to the local feature of facial image and extract, full articulamentum can to each convolutional layer train the local feature extracted to integrate, being trained by each convolutional layer the characteristics of image that extracts to connect is an one-dimensional vector; Output layer is for exporting the classification results of the face figure sample to input.
Be be described based on the degree of deep learning model of convolutional neural networks for described degree of deep learning model below.
1) the model training stage
Based in the degree of deep learning model of convolutional neural networks, each basic unit comprises the node of some linear one dimensional arrangement, between basic unit and the node of basic unit is wherein to be in a kind of full state be connected, and the connection between node has a weight parameter usually.In an initial condition, the weight parameter connected between node is default value, therefore before formal this degree of deep learning model of use carries out face feature vector extraction to facial image to be identified and facial image sample, need to train this degree of deep learning model, to determine the optimal weight parameter of the connection between each node.
Wherein, when training this degree of deep learning model, the facial image sample of predetermined number can be prepared, and these facial image samples are classified; Such as, 50,000 facial image samples can be prepared, then the user belonged to according to these faces figure sample classifies to these 50,000 facial image samples, and each facial image sample of classifying is demarcated according to affiliated user, such as, each classification sorted can be demarcated as Zhang San, Li Si, king five etc. respectively, each user possesses 10 ~ 1000 pictures do not waited, and the facial image sample standard deviation now in each classification belongs to same user.
After the facial image sample classification of the predetermined number prepared completes, now can using this degree of deep learning model as disaggregated model, these facial image samples are input in this degree of deep learning model as training sample and train, and according to the classification results that degree of deep learning model exports, constantly the weight parameter of the connection in this each basic unit of degree of deep learning model between node is adjusted.In continuous adjustment process, this degree of deep learning model is after the training sample based on input is trained, and compared with the classification results that classification results and the user of output demarcate, accuracy will improve gradually.Meanwhile, user can pre-set an accuracy threshold value, in continuous adjustment process, if compared with the classification results that the classification results that this degree of deep learning model exports and user demarcate, after accuracy reaches the accuracy threshold value pre-set, the weight parameter now connected between each base level nodes in this degree of deep learning model is optimal weight parameter, now can think that this degree of deep learning model has been trained complete.
2) model operational phase
For the degree of deep learning model that training is complete, directly can use this degree of deep learning model, facial image to be identified and facial image sample be carried out to the extraction of face feature vector.
When realizing, service end can create a face image sample data storehouse in advance in this locality, each facial image sample standard deviation in this database can when carrying out recognition of face to facial image to be identified, as the object of reference of comparing with this facial image to be identified.
When carrying out face characteristic for the facial image sample in database and extracting, facial image sample in database can be carried out features training successively by service end in multiple convolutional layers that input picture comprises in this degree of deep learning model, after each convolutional layer has all been trained, the face feature vector of proper vector as this facial image sample of full articulamentum output can be extracted.
Due in this degree of deep learning model, full articulamentum can to each convolutional layer train the local feature extracted to integrate, therefore the proper vector exported by full articulamentum is as the face feature vector of this facial image sample, the global characteristics of this face figure sample after being integrated by full articulamentum can be obtained, when using global characteristics to carry out measuring similarity, face recognition result will be more accurate.
Certainly, when realizing, proper vector that any one in multiple convolutional layers of this degree of deep learning model convolutional layer of specifying the exports face feature vector as this facial image sample also can be extracted.By the proper vector that any one convolutional layer of specifying in multiple convolutional layer exports, the local feature of this face figure sample can be obtained.
In the present embodiment, when carrying out face characteristic for facial image to be identified and extracting, service end still can according to identical processing mode, this facial image to be identified is carried out features training successively in multiple convolutional layers that input picture comprises in this degree of deep learning model, after each convolutional layer has all been trained, the face feature vector of proper vector as this facial image to be identified of the convolutional layer output of specifying in full articulamentum or multiple convolutional layer can be extracted.
What deserves to be explained is, service end using the facial image sample in database and facial image to be identified as before input picture carries out features training in this degree of deep learning model, dimension normalization process can be carried out to the facial image sample in database and facial image to be identified centered by eyes, be processed into the image of unified size, compare in the recognition of face stage to facilitate.Wherein, the image size adopted during normalized can be arranged according to the demand of reality; Such as, when being normalized, the image being arranged to 224*224 can be unified.
When after service end to extract facial image sample in database and facial image to be identified face feature vector by this degree of deep learning model, measuring similarity can be carried out based on the face feature vector extracted, calculate face recognition result.
In the present embodiment, when carrying out measuring similarity based on the face feature vector extracted, the vector distance of the face feature vector of the facial image sample in facial image to be identified and database can be utilized to characterize the similarity of the two.
When realizing, service end can calculate the vector distance between the face feature vector of facial image sample in the face feature vector of facial image to be identified and database successively, then the vector distance calculated is converted to corresponding similarity value according to the similarity switching strategy preset.Such as, this similarity switching strategy can be that service end sets up the corresponding relation list of a vector distance and similarity value in advance according to the relation between proper vector and similarity, multiple different similarity grade can be divided into according to the vector distance threshold value preset in this corresponding relation list, and a corresponding similarity value is set for each similarity grade, because the similarity of the vector distance between proper vector usually and between proper vector is inversely proportional to, therefore when vector distance more hour, similarity value is higher, when vector distance is larger, similarity value is lower.Directly just can obtain the similarity value corresponding with the vector distance calculated by this corresponding relation list of inquiry in this way.Wherein, this vector distance can be COS distance, can be also Euclidean distance, not be particularly limited in the present embodiment.
Wherein, service end, before the vector distance of face feature vector calculating described facial image to be identified and described facial image sample, can also carry out dimension-reduction treatment according to the algorithm preset to the face feature vector of the two.Such as, projection matrix training can be carried out by PCA algorithm and LDA algorithm to the face feature vector of the two, dimensionality reduction is carried out to the face feature vector of the two.In this way, the calculated amount of service end can not only be reduced, the redundant information in face feature vector can also be removed, retain the useful information in face feature vector.
After service end similarity value corresponding to the vector distance calculated converts to, can judge whether this similarity value reaches similarity threshold, if when this similarity value reaches similarity threshold, now service end can confirm that this facial image to be identified is identical facial image with this facial image sample, and is exported as recognition result by this facial image sample.Certainly, if when this similarity value does not reach similarity threshold, now service end can confirm that this facial image to be identified is not identical facial image with this facial image sample, now service end can repeat above process, continue the similarity value calculating next face image pattern in this facial image to be identified and database, until find identical facial image, or travel through when whole database does not find facial image to be identified with this identical facial image and stop.
Below by way of a concrete application example, above technical scheme is described in detail.
Refer to Fig. 3, Fig. 3 is a kind of degree of deep learning model based on convolutional Neural net shown in the present embodiment.In this degree of deep learning model, comprise input layer, first volume lamination, volume Two lamination, the 3rd convolutional layer, Volume Four lamination, the 5th convolutional layer, the first full articulamentum, the second full articulamentum and output layer.
In the model training stage, 50,000 facial image samples can be prepared, then the user belonged to according to these faces figure sample carries out manual sort to these 50,000 facial image samples, and each facial image sample of classifying is demarcated according to affiliated user, such as, each classification sorted can be demarcated as Zhang San, Li Si, king five etc. respectively, each user possesses 10 ~ 1000 pictures do not waited, and the facial image sample standard deviation now in each classification belongs to same user.
After the facial image sample classifications of 50,000 complete, can be normalized facial image sample, obtain the image of 224*224.For the facial image sample after normalized, can using this degree of deep learning model as disaggregated model, these facial image samples are input in this degree of deep learning model as training sample and train, and according to the classification results that degree of deep learning model exports, constantly the weight parameter of the connection in this each basic unit of degree of deep learning model between node is adjusted, until the classification results that this degree of deep learning model exports is compared with the classification results manually demarcated, stop when accuracy reaches the accuracy threshold value pre-set.The weight parameter now connected between each base level nodes in this degree of deep learning model is optimal weight parameter, and the training of this degree of deep learning model is complete.
In model operational phase, on the one hand, 50,000 facial image samples can be carried out features training successively by service end in 5 convolutional layers of input picture in this degree of deep learning model, and after training completes, extract the face feature vector of 2*2048 dimensional feature vector as correspondence of full articulamentum output.On the other hand, service end can be treated recognition image and be normalized, obtain the image of 224*224, then features training is carried out successively as in 5 convolutional layers of input picture in this degree of deep learning model, and after training completes, extract the face feature vector of proper vector as this facial image to be identified of the 2*2048 dimension that full articulamentum exports.
In the recognition of face stage, service end can utilize PCA and the LDA algorithm proper vector to the 2*2048 dimension of facial image to be identified and 50,000 facial image samples to carry out dimensionality reduction, obtain the proper vector of one 400 dimension (empirical value), and calculate the COS distance in facial image to be identified and database between these two 400 proper vectors tieed up of facial image sample successively.
For the COS distance calculated, similarity value corresponding to the COS distance calculated can convert to according to the similarity switching strategy preset, and judge whether this similarity value reaches similarity threshold (such as 90%), if when reaching similarity threshold, then can confirm that this facial image to be identified is identical facial image with this facial image sample, and this facial image sample is exported as recognition result.If when this similarity value does not reach similarity threshold, now can repeat above process, continue the similarity value calculating next face image pattern in this facial image to be identified and database, until find identical facial image, or travel through when whole database does not find facial image to be identified with this identical facial image and stop.
What deserves to be explained is, degree of deep learning model illustrated in fig. 3, and the weight parameter, convolution kernel size and the convolution kernel quantity that mark in each layer in this degree of deep learning model are exemplary, and be not used in the restriction disclosure, in the application process of reality, can applicable model be re-created according to the demand of reality or the weight modification marked in above-mentioned model, convolution kernel size and convolution kernel quantity be modified.
In the embodiment above, by extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, and the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted, then according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified, due in the process of recognition of face, degree of deep learning model is utilized automatically from facial image, to extract face characteristic, instead of traditional artificial mode extracting face characteristic from facial image, therefore the degree of accuracy of recognition of face can be improved.
Corresponding with aforementioned face identification method embodiment, the disclosure additionally provides a kind of embodiment of device.
Fig. 4 is the schematic block diagram of a kind of face identification device according to an exemplary embodiment.
As shown in Figure 4, a kind of face identification device 400 according to an exemplary embodiment, comprising: extraction module 401, computing module 402 and identification module 403; Wherein:
Described extraction module 401 is configured to, and extracts the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset;
Described computing module 402 is configured to, and calculates the similarity value of described facial image to be identified and described facial image sample based on the face feature vector extracted;
Described identification module 403 is configured to, and the described similarity value according to calculating carries out recognition of face to described facial image to be identified.
In the embodiment above, by extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, and the similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted, then according to the described similarity value calculated, recognition of face is carried out to described facial image to be identified, due in the process of recognition of face, degree of deep learning model is utilized automatically from facial image, to extract face characteristic, instead of traditional artificial mode extracting face characteristic from facial image, therefore the degree of accuracy of recognition of face can be improved.
Refer to Fig. 5, Fig. 5 is the block diagram of the another kind of device of the disclosure according to an exemplary embodiment, and this embodiment is on aforementioned basis embodiment illustrated in fig. 4, and described device 400 can also comprise training module 405; Wherein:
Described training module 405 is configured to, before extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, facial image sample based on predetermined number is trained described degree of deep learning model, to determine the optimal weight parameter of each node in described degree of deep learning model.
Refer to Fig. 6, Fig. 6 is the block diagram of the another kind of device of the disclosure according to an exemplary embodiment, and this embodiment is on aforementioned basis embodiment illustrated in fig. 4, and described device 400 can also comprise dimensionality reduction module 406; Wherein:
Described dimensionality reduction module 406 is configured to, and before the similarity value calculating described facial image to be identified and described facial image sample based on the face feature vector extracted, carries out dimension-reduction treatment according to preset algorithm to the described face feature vector extracted.
It should be noted that, the structure of the dimensionality reduction module 406 shown in device embodiment shown in above-mentioned Fig. 6 also can be included in the device embodiment of earlier figures 5, does not limit this disclosure.
Refer to Fig. 7, Fig. 7 is the block diagram of the another kind of device of the disclosure according to an exemplary embodiment, and this embodiment is on aforementioned basis embodiment illustrated in fig. 4, and described degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net; Described extraction module 401 can comprise training submodule 401A and extract submodule 401B; Wherein:
Described training submodule 401A is configured to, and described facial image to be identified and described facial image sample are carried out features training successively in multiple basic units that input picture comprises respectively in described degree of deep learning model;
Described extraction submodule 401B is configured to, and after training completes, extracts full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample.
It should be noted that, the structure of the training submodule 401A shown in device embodiment shown in above-mentioned Fig. 7 and extraction submodule 401B also can be included in the device embodiment of earlier figures 5-6, does not limit this disclosure.
Refer to Fig. 8, Fig. 8 is the block diagram of the another kind of device of the disclosure according to an exemplary embodiment, and this embodiment is on aforementioned basis embodiment illustrated in fig. 4, and described computing module 402 can comprise calculating sub module 402A and transform subblock 402B; Wherein:
Described calculating sub module 402A is configured to, the vector distance between the face feature vector calculating described facial image to be identified and described facial image sample; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
Described transform subblock 402B is configured to, and according to the similarity switching strategy preset, the described vector distance calculated is converted to corresponding similarity value.
It should be noted that, the structure of the calculating sub module 402A shown in device embodiment shown in above-mentioned Fig. 8 and transform subblock 402B also can be included in the device embodiment of earlier figures 5-7, does not limit this disclosure.
Refer to Fig. 9, Fig. 9 is the block diagram of the another kind of device of the disclosure according to an exemplary embodiment, and this embodiment is on aforementioned basis embodiment illustrated in fig. 4, and described identification module 403 can comprise and judges submodule 403A and output sub-module 403B; Wherein:
Described judgement submodule 403A is configured to, and judges whether the described similarity value calculated reaches threshold value;
Described output sub-module 403B is configured to, and when described similarity value reaches threshold value, confirms that described facial image to be identified is identical with described facial image sample, and is exported as recognition result by described facial image sample.
It should be noted that, the structure of the judgement submodule 403A shown in device embodiment shown in above-mentioned Fig. 8 and output sub-module 403B also can be included in the device embodiment of earlier figures 5-8, does not limit this disclosure.
In said apparatus, the implementation procedure of the function and efficacy of modules specifically refers to the implementation procedure of corresponding step in said method, does not repeat them here.
For device embodiment, because it corresponds essentially to embodiment of the method, so relevant part illustrates see the part of embodiment of the method.Device embodiment described above is only schematic, the wherein said module illustrated as separating component can or may not be physically separates, parts as module display can be or may not be physical module, namely can be positioned at a place, or also can be distributed on multiple mixed-media network modules mixed-media.Some or all of module wherein can be selected according to the actual needs to realize the object of disclosure scheme.Those of ordinary skill in the art, when not paying creative work, are namely appreciated that and implement.
Accordingly, the disclosure also provides a kind of face identification device, and described device comprises:
Processor;
For the storer of storage of processor executable instruction;
Wherein, described processor is configured to:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
Accordingly, the disclosure also provides a kind of service end, described terminal includes storer, and one or more than one program, one of them or more than one program are stored in storer, and are configured to perform described more than one or one routine package containing the instruction for carrying out following operation by more than one or one processor:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
Accordingly, the disclosure also provides a kind of face identification device, and described device comprises:
Processor;
For the storer of storage of processor executable instruction;
Wherein, described processor is configured to:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
Figure 10 is a kind of block diagram for face identification device 1000 according to an exemplary embodiment.Such as, device 1000 may be provided in a server.With reference to Figure 10, device 1000 comprises processing components 1022, and it comprises one or more processor further, and the memory resource representated by storer 1032, can such as, by the instruction of the execution of processing element 1022, application program for storing.The application program stored in storer 1032 can comprise each module corresponding to one group of instruction one or more.In addition, processing components 1022 is configured to perform instruction, to perform the control method of above-mentioned smart machine.
Device 1000 can also comprise the power management that a power supply module 1026 is configured to actuating unit 1000, and a wired or wireless network interface 1050 is configured to device 1000 to be connected to network, and input and output (I/O) interface 1058.Device 1000 can operate the operating system based on being stored in storer 1032, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or similar.
Those skilled in the art, at consideration instructions and after putting into practice invention disclosed herein, will easily expect other embodiment of the present disclosure.The application is intended to contain any modification of the present disclosure, purposes or adaptations, and these modification, purposes or adaptations are followed general principle of the present disclosure and comprised the undocumented common practise in the art of the disclosure or conventional techniques means.Instructions and embodiment are only regarded as exemplary, and true scope of the present disclosure and spirit are pointed out by claim below.
Should be understood that, the disclosure is not limited to precision architecture described above and illustrated in the accompanying drawings, and can carry out various amendment and change not departing from its scope.The scope of the present disclosure is only limited by appended claim.
Claims (17)
1. a face identification method, is characterized in that, described method comprises:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
2. the method for claim 1, is characterized in that, before the degree of deep learning model that described basis is preset extracts the face feature vector of facial image to be identified and facial image sample, described method also comprises:
Facial image sample based on predetermined number is trained described degree of deep learning model, to determine the optimal weight parameter of the connection in described degree of deep learning model between each node.
3. the method for claim 1, is characterized in that, before the described face feature vector based on extracting calculates the similarity value of described facial image to be identified and described facial image sample, described method comprises:
According to preset algorithm, dimension-reduction treatment is carried out to the described face feature vector extracted.
4. the method for claim 1, is characterized in that, described degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net;
The face feature vector that the degree of deep learning model that described basis is preset extracts facial image to be identified and facial image sample comprises:
Described facial image to be identified and described facial image sample are carried out features training successively in multiple basic units that input picture comprises respectively in described degree of deep learning model;
After training completes, extract full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample.
5. method as claimed in claim 4, is characterized in that, the similarity value that the described face feature vector based on extracting calculates described facial image to be identified and described facial image sample comprises:
Vector distance between the face feature vector calculating described facial image to be identified and described facial image sample; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
According to the similarity switching strategy preset, the described vector distance calculated is converted to corresponding similarity value.
6. method as claimed in claim 5, it is characterized in that, the described similarity value that described basis calculates is carried out recognition of face to described facial image to be identified and is comprised:
Judge whether the described similarity value calculated reaches threshold value;
When described similarity value reaches threshold value, confirm that described facial image to be identified is identical with described facial image sample, and described facial image sample is exported as recognition result.
7. method as claimed in claim 5, it is characterized in that, described vector distance comprises COS distance or Euclidean distance.
8. the method for claim 1, is characterized in that, described facial image to be identified and facial image sample standard deviation have carried out dimension normalization process in advance centered by eyes.
9. a face identification device, is characterized in that, described device comprises:
Extraction module, for extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset;
Computing module, for calculating the similarity value of described facial image to be identified and described facial image sample based on the face feature vector extracted;
Identification module, for carrying out recognition of face according to the described similarity value calculated to described facial image to be identified.
10. method as claimed in claim 9, it is characterized in that, described device also comprises:
Training module, for before extracting the face feature vector of facial image to be identified and facial image sample according to the degree of deep learning model preset, facial image sample based on predetermined number is trained described degree of deep learning model, to determine the optimal weight parameter of each node in described degree of deep learning model.
11. devices as claimed in claim 9, it is characterized in that, described device also comprises:
Dimensionality reduction module, for before the similarity value calculating described facial image to be identified and described facial image sample based on the face feature vector extracted, carries out dimension-reduction treatment according to preset algorithm to the described face feature vector extracted.
12. devices as claimed in claim 9, it is characterized in that, described degree of deep learning model comprises the degree of deep learning model based on convolutional Neural net;
Described extraction module comprises:
Training submodule, for carrying out features training using described facial image to be identified and described facial image sample successively in multiple basic units that input picture comprises respectively in described degree of deep learning model;
Extract submodule, for after training completes, extract full articulamentum in described multiple basic unit or other proper vector of specifying basic unit the to export face feature vector as described facial image to be identified or described facial image sample.
13. devices as claimed in claim 12, it is characterized in that, described computing module comprises:
Calculating sub module, for calculate described facial image to be identified and described facial image sample face feature vector between vector distance; Wherein, described vector distance is for characterizing the similarity between described facial image to be identified and described facial image sample;
Transform subblock, for being converted to corresponding similarity value according to the similarity switching strategy preset by the described vector distance calculated.
14. devices as claimed in claim 13, it is characterized in that, described identification module comprises:
Judge submodule, for judging whether the described similarity value calculated reaches threshold value;
Output sub-module, for when described similarity value reaches threshold value, confirms that described facial image to be identified is identical with described facial image sample, and is exported as recognition result by described facial image sample.
15. devices as claimed in claim 11, it is characterized in that, described vector distance comprises COS distance or Euclidean distance.
16. methods as claimed in claim 9, it is characterized in that, described facial image to be identified and facial image sample standard deviation have carried out dimension normalization process in advance centered by eyes.
17. 1 kinds of face identification devices, is characterized in that, comprising:
Processor;
For the storer of storage of processor executable instruction;
Wherein, described processor is configured to:
The face feature vector of facial image to be identified and facial image sample is extracted according to the degree of deep learning model preset;
The similarity value of described facial image to be identified and described facial image sample is calculated based on the face feature vector extracted;
Described similarity value according to calculating carries out recognition of face to described facial image to be identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510369876.8A CN104899579A (en) | 2015-06-29 | 2015-06-29 | Face recognition method and face recognition device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510369876.8A CN104899579A (en) | 2015-06-29 | 2015-06-29 | Face recognition method and face recognition device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104899579A true CN104899579A (en) | 2015-09-09 |
Family
ID=54032233
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510369876.8A Pending CN104899579A (en) | 2015-06-29 | 2015-06-29 | Face recognition method and face recognition device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104899579A (en) |
Cited By (68)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160361A (en) * | 2015-09-30 | 2015-12-16 | 东软集团股份有限公司 | Image identification method and apparatus |
| CN105335713A (en) * | 2015-10-28 | 2016-02-17 | 小米科技有限责任公司 | Fingerprint identification method and device |
| CN105354560A (en) * | 2015-11-25 | 2016-02-24 | 小米科技有限责任公司 | Fingerprint identification method and device |
| CN105426860A (en) * | 2015-12-01 | 2016-03-23 | 北京天诚盛业科技有限公司 | Human face identification method and apparatus |
| CN105512620A (en) * | 2015-11-30 | 2016-04-20 | 北京天诚盛业科技有限公司 | Convolutional neural network training method and apparatus for face recognition |
| CN105590089A (en) * | 2015-10-22 | 2016-05-18 | 广州视源电子科技股份有限公司 | Face identification method and device |
| CN105608425A (en) * | 2015-12-17 | 2016-05-25 | 小米科技有限责任公司 | Method and device for sorted storage of pictures |
| CN105631403A (en) * | 2015-12-17 | 2016-06-01 | 小米科技有限责任公司 | Method and device for human face recognition |
| CN105718869A (en) * | 2016-01-15 | 2016-06-29 | 网易（杭州）网络有限公司 | Method and device for estimating face score in picture |
| CN105740758A (en) * | 2015-12-31 | 2016-07-06 | 上海极链网络科技有限公司 | Internet video face recognition method based on deep learning |
| CN105740808A (en) * | 2016-01-28 | 2016-07-06 | 北京旷视科技有限公司 | Human face identification method and device |
| CN105868243A (en) * | 2015-12-14 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Information processing method and apparatus |
| CN105930797A (en) * | 2016-04-21 | 2016-09-07 | 腾讯科技（深圳）有限公司 | Face verification method and device |
| CN106022317A (en) * | 2016-06-27 | 2016-10-12 | 北京小米移动软件有限公司 | Face identification method and apparatus |
| CN106202329A (en) * | 2016-07-01 | 2016-12-07 | 北京市商汤科技开发有限公司 | Sample data process, data identification method and device, computer equipment |
| CN106250858A (en) * | 2016-08-05 | 2016-12-21 | 重庆中科云丛科技有限公司 | A kind of recognition methods merging multiple face recognition algorithms and system |
| CN106295707A (en) * | 2016-08-17 | 2017-01-04 | 北京小米移动软件有限公司 | Image-recognizing method and device |
| CN106326867A (en) * | 2016-08-26 | 2017-01-11 | 维沃移动通信有限公司 | Face recognition method and mobile terminal |
| CN106339695A (en) * | 2016-09-20 | 2017-01-18 | 北京小米移动软件有限公司 | Face similarity detection method, device and terminal |
| CN106447625A (en) * | 2016-09-05 | 2017-02-22 | 北京中科奥森数据科技有限公司 | Facial image series-based attribute identification method and device |
| CN106599926A (en) * | 2016-12-20 | 2017-04-26 | 上海寒武纪信息科技有限公司 | Expression picture pushing method and system |
| CN106599855A (en) * | 2016-12-19 | 2017-04-26 | 四川长虹电器股份有限公司 | Softmax-based face recognizing method |
| CN106599837A (en) * | 2016-12-13 | 2017-04-26 | 北京智慧眼科技股份有限公司 | Face identification method and device based on multi-image input |
| CN106611151A (en) * | 2015-10-23 | 2017-05-03 | 杭州海康威视数字技术股份有限公司 | A human face identification method and apparatus |
| CN106713255A (en) * | 2015-11-17 | 2017-05-24 | 腾讯科技（深圳）有限公司 | Use information management method and system |
| CN106791440A (en) * | 2017-01-20 | 2017-05-31 | 奇酷互联网络科技（深圳）有限公司 | Control the method and device of Face datection function |
| CN106777349A (en) * | 2017-01-16 | 2017-05-31 | 广东工业大学 | Face retrieval system and method based on deep learning |
| CN107121436A (en) * | 2017-04-27 | 2017-09-01 | 亚洲硅业（青海）有限公司 | The Intelligent detecting method and identification device of a kind of silicon material quality |
| CN107133576A (en) * | 2017-04-17 | 2017-09-05 | 北京小米移动软件有限公司 | Age of user recognition methods and device |
| CN107203775A (en) * | 2016-03-18 | 2017-09-26 | 阿里巴巴集团控股有限公司 | A kind of method of image classification, device and equipment |
| CN107368770A (en) * | 2016-05-12 | 2017-11-21 | 深圳市维杰乐思科技有限公司 | A kind of frequent customer's automatic identifying method and system |
| CN107563341A (en) * | 2017-09-15 | 2018-01-09 | 赵立峰 | A kind of face identification device and a kind of face identification system |
| CN107578024A (en) * | 2017-09-15 | 2018-01-12 | 赵立峰 | A kind of face identification system |
| CN107679474A (en) * | 2017-09-25 | 2018-02-09 | 北京小米移动软件有限公司 | Face matching process and device |
| CN107808149A (en) * | 2017-11-17 | 2018-03-16 | 腾讯数码（天津）有限公司 | A kind of face information mask method, device and storage medium |
| CN107844784A (en) * | 2017-12-08 | 2018-03-27 | 广东美的智能机器人有限公司 | Face identification method, device, computer equipment and readable storage medium storing program for executing |
| CN107844744A (en) * | 2017-10-09 | 2018-03-27 | 平安科技（深圳）有限公司 | With reference to the face identification method, device and storage medium of depth information |
| CN107886064A (en) * | 2017-11-06 | 2018-04-06 | 安徽大学 | A kind of method that recognition of face scene based on convolutional neural networks adapts to |
| CN108062538A (en) * | 2017-12-29 | 2018-05-22 | 成都智宝大数据科技有限公司 | Face identification method and device |
| CN108090433A (en) * | 2017-12-12 | 2018-05-29 | 厦门集微科技有限公司 | Face identification method and device, storage medium, processor |
| CN108229532A (en) * | 2017-10-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Image-recognizing method, device and electronic equipment |
| CN108429619A (en) * | 2018-01-18 | 2018-08-21 | 北京捷通华声科技股份有限公司 | Identity identifying method and system |
| CN108764031A (en) * | 2018-04-17 | 2018-11-06 | 平安科技（深圳）有限公司 | Identify method, apparatus, computer equipment and the storage medium of face |
| CN108763373A (en) * | 2018-05-17 | 2018-11-06 | 厦门美图之家科技有限公司 | Research on face image retrieval and device |
| CN108875533A (en) * | 2018-01-29 | 2018-11-23 | 北京旷视科技有限公司 | Method, apparatus, system and the computer storage medium of recognition of face |
| CN108875515A (en) * | 2017-12-11 | 2018-11-23 | 北京旷视科技有限公司 | Face identification method, device, system, storage medium and capture machine |
| CN108875834A (en) * | 2018-06-22 | 2018-11-23 | 北京达佳互联信息技术有限公司 | Image clustering method, device, computer equipment and storage medium |
| CN108875767A (en) * | 2017-12-07 | 2018-11-23 | 北京旷视科技有限公司 | Method, apparatus, system and the computer storage medium of image recognition |
| CN109086752A (en) * | 2018-09-30 | 2018-12-25 | 北京达佳互联信息技术有限公司 | Face identification method, device, electronic equipment and storage medium |
| WO2019061662A1 (en) * | 2017-09-30 | 2019-04-04 | 平安科技（深圳）有限公司 | Electronic device, insured domestic animal recognition method and computer readable storage medium |
| CN109583332A (en) * | 2018-11-15 | 2019-04-05 | 北京三快在线科技有限公司 | Face identification method, face identification system, medium and electronic equipment |
| CN109711243A (en) * | 2018-11-01 | 2019-05-03 | 长沙小钴科技有限公司 | A kind of static three-dimensional human face in-vivo detection method based on deep learning |
| CN109727350A (en) * | 2018-12-14 | 2019-05-07 | 深圳壹账通智能科技有限公司 | A kind of Door-access control method and device based on recognition of face |
| CN109741481A (en) * | 2018-12-14 | 2019-05-10 | 深圳壹账通智能科技有限公司 | Body-building duration calculation method, apparatus and body-building equipment based on recognition of face |
| WO2019100436A1 (en) * | 2017-11-22 | 2019-05-31 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face recognition |
| WO2019109768A1 (en) * | 2017-12-04 | 2019-06-13 | 深圳奥比中光科技有限公司 | Task execution method, terminal device and computer readable storage medium |
| CN110070082A (en) * | 2019-04-22 | 2019-07-30 | 苏州科达科技股份有限公司 | Licence plate recognition method, device, equipment and storage medium |
| CN110110189A (en) * | 2018-02-01 | 2019-08-09 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN110210194A (en) * | 2019-04-18 | 2019-09-06 | 深圳壹账通智能科技有限公司 | Electronic contract display methods, device, electronic equipment and storage medium |
| CN110309692A (en) * | 2018-03-27 | 2019-10-08 | 杭州海康威视数字技术股份有限公司 | Face identification method, apparatus and system, model training method and device |
| CN110414315A (en) * | 2019-06-11 | 2019-11-05 | 万翼科技有限公司 | A kind of collaboration supervision method of isomery security system |
| CN110543811A (en) * | 2019-07-15 | 2019-12-06 | 华南理工大学 | non-cooperation type examination person management method and system based on deep learning |
| CN110826390A (en) * | 2019-09-09 | 2020-02-21 | 博云视觉（北京）科技有限公司 | Video data processing method based on face vector characteristics |
| WO2020140377A1 (en) * | 2019-01-04 | 2020-07-09 | 平安科技（深圳）有限公司 | Neural network model training method and apparatus, computer device, and storage medium |
| CN111401291A (en) * | 2020-03-24 | 2020-07-10 | 三一重工股份有限公司 | Stranger identification method and device |
| CN113361654A (en) * | 2021-07-12 | 2021-09-07 | 广州天鹏计算机科技有限公司 | Image identification method and system based on machine learning |
| WO2021232985A1 (en) * | 2020-05-22 | 2021-11-25 | 腾讯科技（深圳）有限公司 | Facial recognition method and apparatus, computer device, and storage medium |
| US11455489B2 (en) * | 2018-06-13 | 2022-09-27 | Canon Kabushiki Kaisha | Device that updates recognition model and method of updating recognition model |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070122009A1 (en) * | 2005-11-26 | 2007-05-31 | Hyung Keun Jee | Face recognition method and apparatus |
| CN103824054A (en) * | 2014-02-17 | 2014-05-28 | 北京旷视科技有限公司 | Cascaded depth neural network-based face attribute recognition method |
| CN104077563A (en) * | 2014-05-30 | 2014-10-01 | 小米科技有限责任公司 | Human face recognition method and device |
| CN104463172A (en) * | 2014-12-09 | 2015-03-25 | 中国科学院重庆绿色智能技术研究院 | Face feature extraction method based on face feature point shape drive depth model |
| CN104573681A (en) * | 2015-02-11 | 2015-04-29 | 成都果豆数字娱乐有限公司 | Face recognition method |
| CN104657709A (en) * | 2015-02-05 | 2015-05-27 | 小米科技有限责任公司 | Face image identification method and device as well as server |
| CN104680119A (en) * | 2013-11-29 | 2015-06-03 | 华为技术有限公司 | Image identity recognition method, related device and identity recognition system |
-
2015
- 2015-06-29 CN CN201510369876.8A patent/CN104899579A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070122009A1 (en) * | 2005-11-26 | 2007-05-31 | Hyung Keun Jee | Face recognition method and apparatus |
| US7873189B2 (en) * | 2005-11-26 | 2011-01-18 | Electronics And Telecommunications Research Institute | Face recognition by dividing an image and evaluating a similarity vector with a support vector machine |
| CN104680119A (en) * | 2013-11-29 | 2015-06-03 | 华为技术有限公司 | Image identity recognition method, related device and identity recognition system |
| CN103824054A (en) * | 2014-02-17 | 2014-05-28 | 北京旷视科技有限公司 | Cascaded depth neural network-based face attribute recognition method |
| CN104077563A (en) * | 2014-05-30 | 2014-10-01 | 小米科技有限责任公司 | Human face recognition method and device |
| CN104463172A (en) * | 2014-12-09 | 2015-03-25 | 中国科学院重庆绿色智能技术研究院 | Face feature extraction method based on face feature point shape drive depth model |
| CN104657709A (en) * | 2015-02-05 | 2015-05-27 | 小米科技有限责任公司 | Face image identification method and device as well as server |
| CN104573681A (en) * | 2015-02-11 | 2015-04-29 | 成都果豆数字娱乐有限公司 | Face recognition method |
Cited By (95)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160361A (en) * | 2015-09-30 | 2015-12-16 | 东软集团股份有限公司 | Image identification method and apparatus |
| CN105590089A (en) * | 2015-10-22 | 2016-05-18 | 广州视源电子科技股份有限公司 | Face identification method and device |
| CN106611151A (en) * | 2015-10-23 | 2017-05-03 | 杭州海康威视数字技术股份有限公司 | A human face identification method and apparatus |
| CN105335713A (en) * | 2015-10-28 | 2016-02-17 | 小米科技有限责任公司 | Fingerprint identification method and device |
| CN106713255B (en) * | 2015-11-17 | 2020-07-24 | 腾讯科技（深圳）有限公司 | User information management method and system |
| CN106713255A (en) * | 2015-11-17 | 2017-05-24 | 腾讯科技（深圳）有限公司 | Use information management method and system |
| CN105354560A (en) * | 2015-11-25 | 2016-02-24 | 小米科技有限责任公司 | Fingerprint identification method and device |
| CN105512620A (en) * | 2015-11-30 | 2016-04-20 | 北京天诚盛业科技有限公司 | Convolutional neural network training method and apparatus for face recognition |
| CN105512620B (en) * | 2015-11-30 | 2019-07-26 | 北京眼神智能科技有限公司 | The training method and device of convolutional neural networks for recognition of face |
| CN105426860A (en) * | 2015-12-01 | 2016-03-23 | 北京天诚盛业科技有限公司 | Human face identification method and apparatus |
| CN105426860B (en) * | 2015-12-01 | 2019-09-27 | 北京眼神智能科技有限公司 | The method and apparatus of recognition of face |
| CN105868243A (en) * | 2015-12-14 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Information processing method and apparatus |
| CN105608425A (en) * | 2015-12-17 | 2016-05-25 | 小米科技有限责任公司 | Method and device for sorted storage of pictures |
| CN105631403A (en) * | 2015-12-17 | 2016-06-01 | 小米科技有限责任公司 | Method and device for human face recognition |
| CN105608425B (en) * | 2015-12-17 | 2019-02-15 | 小米科技有限责任公司 | The method and device of classification storage is carried out to photo |
| CN105631403B (en) * | 2015-12-17 | 2019-02-12 | 小米科技有限责任公司 | Face identification method and device |
| CN105740758A (en) * | 2015-12-31 | 2016-07-06 | 上海极链网络科技有限公司 | Internet video face recognition method based on deep learning |
| CN105718869A (en) * | 2016-01-15 | 2016-06-29 | 网易（杭州）网络有限公司 | Method and device for estimating face score in picture |
| CN105718869B (en) * | 2016-01-15 | 2019-07-02 | 网易（杭州）网络有限公司 | The method and apparatus of face face value in a kind of assessment picture |
| CN105740808A (en) * | 2016-01-28 | 2016-07-06 | 北京旷视科技有限公司 | Human face identification method and device |
| CN105740808B (en) * | 2016-01-28 | 2019-08-09 | 北京旷视科技有限公司 | Face identification method and device |
| CN107203775B (en) * | 2016-03-18 | 2021-07-27 | 斑马智行网络(香港)有限公司 | Image classification method, device and equipment |
| CN107203775A (en) * | 2016-03-18 | 2017-09-26 | 阿里巴巴集团控股有限公司 | A kind of method of image classification, device and equipment |
| CN105930797B (en) * | 2016-04-21 | 2019-03-26 | 腾讯科技（深圳）有限公司 | A kind of face verification method and device |
| CN105930797A (en) * | 2016-04-21 | 2016-09-07 | 腾讯科技（深圳）有限公司 | Face verification method and device |
| CN107368770A (en) * | 2016-05-12 | 2017-11-21 | 深圳市维杰乐思科技有限公司 | A kind of frequent customer's automatic identifying method and system |
| CN106022317A (en) * | 2016-06-27 | 2016-10-12 | 北京小米移动软件有限公司 | Face identification method and apparatus |
| CN106202329A (en) * | 2016-07-01 | 2016-12-07 | 北京市商汤科技开发有限公司 | Sample data process, data identification method and device, computer equipment |
| WO2018001384A1 (en) * | 2016-07-01 | 2018-01-04 | 北京市商汤科技开发有限公司 | Data processing, data recognition methods and devices, and computer equipment |
| CN106250858A (en) * | 2016-08-05 | 2016-12-21 | 重庆中科云丛科技有限公司 | A kind of recognition methods merging multiple face recognition algorithms and system |
| CN106250858B (en) * | 2016-08-05 | 2021-08-13 | 重庆中科云从科技有限公司 | Recognition method and system fusing multiple face recognition algorithms |
| CN106295707A (en) * | 2016-08-17 | 2017-01-04 | 北京小米移动软件有限公司 | Image-recognizing method and device |
| CN106295707B (en) * | 2016-08-17 | 2019-07-02 | 北京小米移动软件有限公司 | Image-recognizing method and device |
| CN106326867A (en) * | 2016-08-26 | 2017-01-11 | 维沃移动通信有限公司 | Face recognition method and mobile terminal |
| CN106326867B (en) * | 2016-08-26 | 2019-06-07 | 维沃移动通信有限公司 | A kind of method and mobile terminal of recognition of face |
| CN106447625A (en) * | 2016-09-05 | 2017-02-22 | 北京中科奥森数据科技有限公司 | Facial image series-based attribute identification method and device |
| CN106339695A (en) * | 2016-09-20 | 2017-01-18 | 北京小米移动软件有限公司 | Face similarity detection method, device and terminal |
| CN106339695B (en) * | 2016-09-20 | 2019-11-15 | 北京小米移动软件有限公司 | Face similarity detection method, device and terminal |
| CN106599837A (en) * | 2016-12-13 | 2017-04-26 | 北京智慧眼科技股份有限公司 | Face identification method and device based on multi-image input |
| CN106599855A (en) * | 2016-12-19 | 2017-04-26 | 四川长虹电器股份有限公司 | Softmax-based face recognizing method |
| CN106599926A (en) * | 2016-12-20 | 2017-04-26 | 上海寒武纪信息科技有限公司 | Expression picture pushing method and system |
| CN106777349A (en) * | 2017-01-16 | 2017-05-31 | 广东工业大学 | Face retrieval system and method based on deep learning |
| CN106791440A (en) * | 2017-01-20 | 2017-05-31 | 奇酷互联网络科技（深圳）有限公司 | Control the method and device of Face datection function |
| CN107133576A (en) * | 2017-04-17 | 2017-09-05 | 北京小米移动软件有限公司 | Age of user recognition methods and device |
| CN107121436B (en) * | 2017-04-27 | 2018-06-01 | 亚洲硅业（青海）有限公司 | The Intelligent detecting method and identification device of a kind of silicon material quality |
| CN107121436A (en) * | 2017-04-27 | 2017-09-01 | 亚洲硅业（青海）有限公司 | The Intelligent detecting method and identification device of a kind of silicon material quality |
| CN107563341A (en) * | 2017-09-15 | 2018-01-09 | 赵立峰 | A kind of face identification device and a kind of face identification system |
| CN107578024A (en) * | 2017-09-15 | 2018-01-12 | 赵立峰 | A kind of face identification system |
| CN107563341B (en) * | 2017-09-15 | 2023-03-24 | 威海元程信息科技有限公司 | Face recognition device and face recognition system |
| CN107679474A (en) * | 2017-09-25 | 2018-02-09 | 北京小米移动软件有限公司 | Face matching process and device |
| WO2019061662A1 (en) * | 2017-09-30 | 2019-04-04 | 平安科技（深圳）有限公司 | Electronic device, insured domestic animal recognition method and computer readable storage medium |
| CN107844744A (en) * | 2017-10-09 | 2018-03-27 | 平安科技（深圳）有限公司 | With reference to the face identification method, device and storage medium of depth information |
| CN108229532A (en) * | 2017-10-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Image-recognizing method, device and electronic equipment |
| CN108229532B (en) * | 2017-10-30 | 2021-02-12 | 北京市商汤科技开发有限公司 | Image recognition method and device and electronic equipment |
| CN107886064A (en) * | 2017-11-06 | 2018-04-06 | 安徽大学 | A kind of method that recognition of face scene based on convolutional neural networks adapts to |
| CN107886064B (en) * | 2017-11-06 | 2021-10-22 | 安徽大学 | Face recognition scene adaptation method based on convolutional neural network |
| CN107808149A (en) * | 2017-11-17 | 2018-03-16 | 腾讯数码（天津）有限公司 | A kind of face information mask method, device and storage medium |
| WO2019100436A1 (en) * | 2017-11-22 | 2019-05-31 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face recognition |
| US11651229B2 (en) | 2017-11-22 | 2023-05-16 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face recognition |
| WO2019109768A1 (en) * | 2017-12-04 | 2019-06-13 | 深圳奥比中光科技有限公司 | Task execution method, terminal device and computer readable storage medium |
| CN108875767A (en) * | 2017-12-07 | 2018-11-23 | 北京旷视科技有限公司 | Method, apparatus, system and the computer storage medium of image recognition |
| CN107844784A (en) * | 2017-12-08 | 2018-03-27 | 广东美的智能机器人有限公司 | Face identification method, device, computer equipment and readable storage medium storing program for executing |
| CN108875515A (en) * | 2017-12-11 | 2018-11-23 | 北京旷视科技有限公司 | Face identification method, device, system, storage medium and capture machine |
| CN108090433A (en) * | 2017-12-12 | 2018-05-29 | 厦门集微科技有限公司 | Face identification method and device, storage medium, processor |
| CN108062538A (en) * | 2017-12-29 | 2018-05-22 | 成都智宝大数据科技有限公司 | Face identification method and device |
| CN108429619A (en) * | 2018-01-18 | 2018-08-21 | 北京捷通华声科技股份有限公司 | Identity identifying method and system |
| CN108875533A (en) * | 2018-01-29 | 2018-11-23 | 北京旷视科技有限公司 | Method, apparatus, system and the computer storage medium of recognition of face |
| US10832032B2 (en) | 2018-01-29 | 2020-11-10 | Beijing Kuangshi Technology Co., Ltd. | Facial recognition method, facial recognition system, and non-transitory recording medium |
| CN110110189A (en) * | 2018-02-01 | 2019-08-09 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN110309692A (en) * | 2018-03-27 | 2019-10-08 | 杭州海康威视数字技术股份有限公司 | Face identification method, apparatus and system, model training method and device |
| CN108764031A (en) * | 2018-04-17 | 2018-11-06 | 平安科技（深圳）有限公司 | Identify method, apparatus, computer equipment and the storage medium of face |
| CN108764031B (en) * | 2018-04-17 | 2023-04-18 | 平安科技（深圳）有限公司 | Method, device, computer equipment and storage medium for recognizing human face |
| CN108763373A (en) * | 2018-05-17 | 2018-11-06 | 厦门美图之家科技有限公司 | Research on face image retrieval and device |
| US11455489B2 (en) * | 2018-06-13 | 2022-09-27 | Canon Kabushiki Kaisha | Device that updates recognition model and method of updating recognition model |
| CN108875834B (en) * | 2018-06-22 | 2019-08-20 | 北京达佳互联信息技术有限公司 | Image clustering method, device, computer equipment and storage medium |
| CN108875834A (en) * | 2018-06-22 | 2018-11-23 | 北京达佳互联信息技术有限公司 | Image clustering method, device, computer equipment and storage medium |
| CN109086752A (en) * | 2018-09-30 | 2018-12-25 | 北京达佳互联信息技术有限公司 | Face identification method, device, electronic equipment and storage medium |
| CN109711243A (en) * | 2018-11-01 | 2019-05-03 | 长沙小钴科技有限公司 | A kind of static three-dimensional human face in-vivo detection method based on deep learning |
| CN109583332A (en) * | 2018-11-15 | 2019-04-05 | 北京三快在线科技有限公司 | Face identification method, face identification system, medium and electronic equipment |
| CN109583332B (en) * | 2018-11-15 | 2021-07-27 | 北京三快在线科技有限公司 | Face recognition method, face recognition system, medium, and electronic device |
| CN109727350A (en) * | 2018-12-14 | 2019-05-07 | 深圳壹账通智能科技有限公司 | A kind of Door-access control method and device based on recognition of face |
| CN109741481A (en) * | 2018-12-14 | 2019-05-10 | 深圳壹账通智能科技有限公司 | Body-building duration calculation method, apparatus and body-building equipment based on recognition of face |
| WO2020140377A1 (en) * | 2019-01-04 | 2020-07-09 | 平安科技（深圳）有限公司 | Neural network model training method and apparatus, computer device, and storage medium |
| CN110210194A (en) * | 2019-04-18 | 2019-09-06 | 深圳壹账通智能科技有限公司 | Electronic contract display methods, device, electronic equipment and storage medium |
| CN110070082A (en) * | 2019-04-22 | 2019-07-30 | 苏州科达科技股份有限公司 | Licence plate recognition method, device, equipment and storage medium |
| CN110414315A (en) * | 2019-06-11 | 2019-11-05 | 万翼科技有限公司 | A kind of collaboration supervision method of isomery security system |
| CN110543811B (en) * | 2019-07-15 | 2024-03-08 | 华南理工大学 | Deep learning-based non-cooperative examination personnel management method and system |
| CN110543811A (en) * | 2019-07-15 | 2019-12-06 | 华南理工大学 | non-cooperation type examination person management method and system based on deep learning |
| CN110826390A (en) * | 2019-09-09 | 2020-02-21 | 博云视觉（北京）科技有限公司 | Video data processing method based on face vector characteristics |
| CN110826390B (en) * | 2019-09-09 | 2023-09-08 | 博云视觉（北京）科技有限公司 | Video data processing method based on face vector characteristics |
| CN111401291B (en) * | 2020-03-24 | 2023-07-14 | 盛景智能科技(嘉兴)有限公司 | Stranger identification method and device |
| CN111401291A (en) * | 2020-03-24 | 2020-07-10 | 三一重工股份有限公司 | Stranger identification method and device |
| WO2021232985A1 (en) * | 2020-05-22 | 2021-11-25 | 腾讯科技（深圳）有限公司 | Facial recognition method and apparatus, computer device, and storage medium |
| US11816880B2 (en) | 2020-05-22 | 2023-11-14 | Tencent Technology (Shenzhen) Company Limited | Face recognition method and apparatus, computer device, and storage medium |
| CN113361654A (en) * | 2021-07-12 | 2021-09-07 | 广州天鹏计算机科技有限公司 | Image identification method and system based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104899579A (en) | Face recognition method and face recognition device | |
| WO2019120115A1 (en) | Facial recognition method, apparatus, and computer apparatus | |
| CN106022317A (en) | Face identification method and apparatus | |
| CN110309840A (en) | Risk trade recognition methods, device, server and storage medium | |
| CN106294344B (en) | Video retrieval method and device | |
| CN108229298A (en) | The training of neural network and face identification method and device, equipment, storage medium | |
| CN109426781A (en) | Construction method, face identification method, device and the equipment of face recognition database | |
| CN110765882B (en) | Video tag determination method, device, server and storage medium | |
| CN105981041A (en) | Facial landmark localization using coarse-to-fine cascaded neural networks | |
| CN106203242A (en) | A kind of similar image recognition methods and equipment | |
| CN109284675A (en) | A kind of recognition methods of user, device and equipment | |
| CN105740808B (en) | Face identification method and device | |
| CN107679457A (en) | User identity method of calibration and device | |
| CN108197669B (en) | Feature training method and device of convolutional neural network | |
| CN110163631A (en) | The accurate method of servicing of client, device, computer equipment and storage medium | |
| CN108492301A (en) | A kind of Scene Segmentation, terminal and storage medium | |
| CN107871103A (en) | Face authentication method and device | |
| CN115984930A (en) | Micro expression recognition method and device and micro expression recognition model training method | |
| CN114511083A (en) | Model training method and device, storage medium and electronic device | |
| CN113762049B (en) | Content identification method, content identification device, storage medium and terminal equipment | |
| CN104978569A (en) | Sparse representation based incremental face recognition method | |
| CN107784482A (en) | Recruitment methods, electronic installation and readable storage medium storing program for executing | |
| CN113033507B (en) | Scene recognition method and device, computer equipment and storage medium | |
| CN114547365A (en) | Image retrieval method and device | |
| CN113254687A (en) | Image retrieval and image quantification model training method, device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into 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: 20150909 |