CN110378235A - A kind of fuzzy facial image recognition method, device and terminal device - Google Patents
A kind of fuzzy facial image recognition method, device and terminal device Download PDFInfo
- Publication number
- CN110378235A CN110378235A CN201910536758.XA CN201910536758A CN110378235A CN 110378235 A CN110378235 A CN 110378235A CN 201910536758 A CN201910536758 A CN 201910536758A CN 110378235 A CN110378235 A CN 110378235A
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- image
- face
- processed
- human face
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000001815 facial effect Effects 0.000 title claims abstract description 143
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000005520 cutting process Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 16
- 230000033001 locomotion Effects 0.000 claims description 12
- 238000003745 diagnosis Methods 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 8
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 230000003252 repetitive effect Effects 0.000 claims description 2
- 238000013145 classification model Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000007689 inspection Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 239000012141 concentrate Substances 0.000 description 3
- 230000009193 crawling Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Multimedia (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of fuzzy facial image recognition method, device and terminal devices, suitable for technical field of data processing, this method comprises: carrying out facial contour cutting to image to be processed and adjusting to preset image sizes, corresponding human face region image to be processed is obtained；Human face region image to be processed is input in preparatory trained Fuzzy Identification Model and is handled, identify whether human face region image to be processed is blurred picture, if human face region image to be processed is blurred picture, recognition of face is then carried out to image to be processed based on preset human face recognition model, judges that image to be processed includes the probability of face；If the probability that image to be processed includes face is less than default face probability threshold value, determine that image to be processed is fuzzy facial image.The embodiment of the present invention greatly improves finally to the compatibility and accuracy rate of different types of fuzzy facial image identification.
Description
Technical field
The invention belongs to technical field of data processing more particularly to fuzzy facial image recognition methods and terminal device.
Background technique
Acquisition, transmission and the preservation to facial image during, due to the movement of capture apparatus or user, environment
The factors such as the configuration of light and capture apparatus influence, and it is fuzzy to will cause finally obtained facial image, and are using these moulds
Paste facial image carries out in next step in application, the effect of practical application can greatly be influenced, such as the sample as human face identification work-attendance checking
When notebook data, the matched reliability of face can be greatly reduced, it is therefore desirable to the method that one kind can identify fuzzy facial image,
To realize that the identification to fuzzy facial image is screened.
The method that some fuzzy image recognitions exist in the prior art, but both for the lower blurred picture of resolution ratio into
Row processing identification, on the one hand as described above, there are many possible factor for causing facial image fuzzy, therefore may deposit in practical application
Fuzzy facial image type also there are many, such as motion blur caused by object of which movement, light it is bad caused by light mould
It pastes and since the lower caused resolution ratio of shooting resolution ratio is fuzzy, locates since the prior art is only capable of obscuring resolution ratio
Reason, it is unsatisfactory so as to cause final actual recognition effect for other kinds of fuzzy image recognition effect and unfriendly,
On the other hand, since these fuzzy image recognitions do not combine face itself to be designed, so that being obscured to face
The influence in non-face region in image is highly prone to when identification, to cause the situation of identification inaccuracy, therefore, existing skill
Art is badly in need of one kind and is adapted to various types of fuzzy facial image recognition method.
Summary of the invention
In view of this, the embodiment of the invention provides a kind of fuzzy facial image recognition method and terminal device, to solve
It can not be adapted to various types of fuzzy facial image identification in the prior art, be compatible with so that being identified to fuzzy facial image
Property and the lower problem of accuracy rate.
The first aspect of the embodiment of the present invention provides a kind of fuzzy facial image recognition method, comprising:
Facial contour cutting is carried out to image to be processed and is adjusted to preset image sizes, corresponding face to be processed is obtained
Area image；
The human face region image to be processed is input in preparatory trained Fuzzy Identification Model and is handled, is identified
Whether the human face region image to be processed is blurred picture, wherein Fuzzy Identification Model is to be in advance based on clear human face region
The identification model that image pattern and the fuzzy human face region image pattern of multi-Fuzzy type are trained, schemes for identification
It seem no for blurred picture；
If the human face region image to be processed is blurred picture, based on preset human face recognition model to described wait locate
It manages image and carries out recognition of face, judge that the image to be processed includes the probability of face；
If the probability that the image to be processed includes face is less than default face probability threshold value, the figure to be processed is determined
As being fuzzy facial image.
The second aspect of the embodiment of the present invention provides a kind of fuzzy facial image identification device, comprising:
Face cuts module, for carrying out facial contour cutting to image to be processed and adjusting to preset image sizes, obtains
To corresponding human face region image to be processed；
Fuzzy image recognition module, for the human face region image to be processed to be input to preparatory trained fuzzy knowledge
It is handled in other model, identifies whether the human face region image to be processed is blurred picture, wherein Fuzzy Identification Model is
The fuzzy human face region image pattern for being in advance based on clear human face region image pattern and multi-Fuzzy type is trained to obtain
Identification model, whether image is blurred picture for identification；
Face recognition module is based on preset face if being blurred picture for the human face region image to be processed
Identification model carries out recognition of face to the image to be processed, judges that the image to be processed includes the probability of face；
Fuzzy face recognition module, if the probability for the image to be processed to include face is less than default face probability threshold
Value then determines that the image to be processed is fuzzy facial image.
The third aspect of the embodiment of the present invention provides a kind of terminal device, and the terminal device includes memory, processing
Device, the computer program that can be run on the processor is stored on the memory, and the processor executes the calculating
The step of fuzzy facial image recognition method as described above is realized when machine program.
The fourth aspect of the embodiment of the present invention provides a kind of computer readable storage medium, comprising: is stored with computer
Program, which is characterized in that fuzzy facial image identification side as described above is realized when the computer program is executed by processor
The step of method.
Existing beneficial effect is the embodiment of the present invention compared with prior art: on the one hand by advance to image to be processed
It carries out facial contour to cut with picture size adjustment, had not only remained crucial face characteristic, but also to sentence to what facial image obscured
It is disconnected to concentrate on face itself, the clarity interference in non-face region is reduced, on the other hand, then by based on practical clear people
The Fuzzy Identification Model of face area image sample and the training building of the fuzzy human face region image pattern of multi-Fuzzy type, is treated
It handles human face region image and carries out fuzzy diagnosis, realize and the compatibility of a variety of face vague category identifiers is accurately identified, accurately sentence
Whether the human face region to be processed that breaks obscures, meanwhile, it may be deposited in facial contour identification and Fuzzy Identification Model identification in order to prevent
Interference cause identification error to increase, the embodiment of the present invention also further to be judged as the image to be processed of blurred picture into
Row whether include face secondary verification, and in comprising the lesser situation of face probability, i.e., various fuzzy factors lead to face
In the lower situation of recognition accuracy, just finally determine that image to be processed is fuzzy facial image, to greatly improve most
Eventually to the compatibility and accuracy rate of different types of fuzzy facial image identification.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only of the invention some
Embodiment for those of ordinary skill in the art without any creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is the implementation process schematic diagram for the fuzzy facial image recognition method that the embodiment of the present invention one provides；
Fig. 2 is the implementation process schematic diagram of fuzzy facial image recognition method provided by Embodiment 2 of the present invention；
Fig. 3 is the implementation process schematic diagram for the fuzzy facial image recognition method that the embodiment of the present invention three provides；
Fig. 4 is the implementation process schematic diagram for the fuzzy facial image recognition method that the embodiment of the present invention four provides；
Fig. 5 is the structural schematic diagram for the fuzzy facial image identification device that the embodiment of the present invention five provides；
Fig. 6 is the schematic diagram for the terminal device that the embodiment of the present invention six provides.
Specific embodiment
In being described below, for illustration and not for limitation, the tool of such as particular system structure, technology etc is proposed
Body details, to understand thoroughly the embodiment of the present invention.However, it will be clear to one skilled in the art that there is no these specific
The present invention also may be implemented in the other embodiments of details.In other situations, it omits to well-known system, device, electricity
The detailed description of road and method, in case unnecessary details interferes description of the invention.
In order to illustrate technical solutions according to the invention, the following is a description of specific embodiments.
To facilitate the understanding of the present invention, first the embodiment of the present invention is briefly described herein, due to being made in actual conditions
There are many possible factor obscured at facial image, also have so as to cause fuzzy facial image type that may be present in practical application
It is a variety of, for example, motion blur caused by object of which movement, light it is bad caused by light it is fuzzy and since shooting resolution ratio is lower
Caused resolution ratio is fuzzy, although in the prior art there are also fuzzy image recognition method, it is lower both for resolution ratio
Blurred picture identification, and be all not bound with face itself and be designed, so that the prior art is fuzzy to different type
The accuracy of facial image identification is difficult to effectively be ensured.
In order to realize the compatible identification and the raising recognition accuracy that obscure facial image to different type, present invention implementation
Various types of fuzzy facial image sample can be acquired in advance first in example, and carries out fuzzy diagnosis based on these samples
The training of model, to obtain to carry out the Fuzzy Identification Model of fuzzy image recognition, on this basis, on the one hand by pre-
Facial contour cutting first is carried out to image to be processed and picture size adjusts, had not only remained crucial face characteristic, but also make to people
The fuzzy judgement of face image can concentrate on face itself, reduce the clarity interference in non-face region, on the other hand, then pass through
Fuzzy diagnosis is carried out to human face region image to be processed based on trained Fuzzy Identification Model, is realized fuzzy to a variety of faces
The compatibility of type accurately identifies, and whether accurate judgement human face region to be processed obscures, while facial contour identifies in order to prevent
Identification error is caused to increase with interference that may be present in Fuzzy Identification Model identification, the embodiment of the present invention is also further to sentencing
Break and be made whether the secondary verification comprising face for the image to be processed of blurred picture, and is including the lesser situation of face probability
Under, i.e., various fuzzy factors cause in the lower situation of face recognition accuracy rate, just finally determine that image to be processed is fuzzy people
Face image, to greatly improve finally to the compatibility and accuracy rate of different types of fuzzy facial image identification.It is described in detail
It is as follows:
Fig. 1 shows the implementation flow chart of the fuzzy facial image recognition method of the offer of the embodiment of the present invention one, is described in detail such as
Under:
S101 carries out facial contour cutting to image to be processed and adjusts to preset image sizes, obtains corresponding wait locate
Manage human face region image.
In actual conditions, even specifically for the image that face is shot, such as certificate photo and take pictures certainly,
Wherein also can be comprising a large amount of non-face regions, and traditional fuzzy recognition methods is all that fuzzy knowledge is carried out directly against whole image
Not, the fog-level in non-face region will greatly influence final fuzzy Judgment at this time, such as during self-timer often
Using background blurring technology is arrived, this makes background area non-face in finally obtained image extremely fuzzy, and the prior art exists
Be easy to when fuzzy diagnosis by whole image recognition being blurred picture.Therefore, fuzzy facial image is identified in order to improve
Accuracy, while retaining crucial face characteristic, the embodiment of the present invention can carry out facial contour cutting to image to be processed first,
It identifies the human face region in image to be processed, and is cut and be independent human face region image to be processed.Simultaneously as
There is some difference for the human face region picture size meeting for including in different images to be processed, is not easy to subsequent Fuzzy Identification Model
Processing identification, therefore the embodiment of the present invention can be unified the human face region extracted carrying out size scaling, adjust to preset figure
As size, wherein specific picture size can be set according to actual needs by technical staff, it is preferable that can be set to
112 × 112 pixels.
It should be explanatorily to be processed since the embodiment of the present invention inherently needs to identify fuzzy facial image
The human face region for including in image be also particularly likely that it is fuzzy so that being inherently difficult to accurately to the identification of human face region
Change, if the requirement for carrying out recognition of face is excessively high to most likely result in can not identify to some fuzzy human face regions, thus can not
Facial contour cutting normally is carried out to image to be processed, therefore the embodiment of the present invention should be adopted when carrying out facial contour cutting
Recognition of face lookup is carried out with the lower face identification method of some precise requirements and is cut, including but not limited to such as face wheel
Wide location algorithm carries out facial contour lookup, and the area image for meeting facial contour is identified as human face region and is cut out
Certainly, or by the recognition of face judgment threshold in some traditional face recognition algorithms reduce, to improve to fuzzy face
Recognition capability.
As an embodiment of the present invention, the interference of other objects in order to prevent, so that being obtained when recognition of face multiple
Human face region, such as similar facial contour object and shooting when other faces of non-targeted face for accidentally photographing, this hair
Bright embodiment only will can wherein meet recognition of face requirement and the most region recognition of pixel when carrying out human face region identification
For human face region, and carry out subsequent cutting.
Human face region image to be processed is input in preparatory trained Fuzzy Identification Model and handles, knows by S102
Whether human face region image not to be processed is blurred picture, wherein Fuzzy Identification Model is to be in advance based on clear human face region figure
The identification model that the fuzzy human face region image pattern of decent and multi-Fuzzy type is trained, for identification image
It whether is blurred picture.
In embodiments of the present invention, it in order to realize the compatible identification to different types of fuzzy people's image, can acquire in advance
A large amount of various types of fuzzy facial image samples, and with carry out model training obtain corresponding Fuzzy Identification Model,
In, the version of Fuzzy Identification Model includes but is not limited to such as two disaggregated models and neural network model, specifically by technology people
Member selectes according to actual needs, not limits herein, while specific model training method can also have technical staff's sets itself,
Or see also the embodiment of the present invention two and other related embodiment contents of the present invention.
On the basis of training Fuzzy Identification Model in advance, the embodiment of the present invention can will cut obtained face to be processed
Area image is input in Fuzzy Identification Model, is carried out processing using model and is identified whether human face region to be processed is fuzzy graph
Picture.
S103, if human face region image to be processed is blurred picture, based on preset human face recognition model to be processed
Image carries out recognition of face, judges that image to be processed includes the probability of face.
S104 determines image to be processed if the probability that image to be processed includes face is less than default face probability threshold value
To obscure facial image.
Since whether fuzzy itself is a opposite concept, (obscuring which kind of degree just be can be determined that as fuzzy graph
Picture), the label that technical staff marks image pattern is to rely on when carrying out Fuzzy Identification Model training to carry out
Category division, therefore the Fuzzy Identification Model that different technologies personnel training obtains, can in the standard for carrying out fuzzy diagnosis judgement
It can have a certain difference, while in actual conditions, with the continuous development of various face processing algorithms, to fog-level phase
Great promotion is also obtained to the identifying processing ability for not being very high facial image, if at this time by the lower people of fog-level
Face image is also determined as fuzzy facial image, can to propose the face image data being collected it is higher upgrade demand,
To bring biggish interference to practical application man face image acquiring, so that the efficiency of acquisition reduces, therefore, in practical application
In, the judgement demand whether facial image obscures should be more biased towards and picked out in by the higher facial image of fog-level.
These above-mentioned practical application requests are based on, human face region image recognition to be processed is being blurred picture by the embodiment of the present invention
When, it can't be directly determined as fuzzy facial image, but secondary verification can be carried out to it, to guarantee final recognition result
It is able to satisfy the demand of practical application, specifically:
In view of no matter which kind of fuzzy excitation factor finally can all cause human face region itself to thicken, therefore the present invention
Embodiment can carry out recognition of face and obtain in it in secondary check to the human face region image to be processed of blurred picture is determined as
Probability comprising face is less than default face probability threshold value situation for probability, that is, illustrates in the human face region image to be processed
Face fog-level is higher, it is difficult to normally identify face therein, therefore the embodiment of the present invention can directly determine in it at this time
Face is fuzzy, i.e., image to be processed is fuzzy facial image.Wherein, specifically used human face recognition model not limits herein,
But unlike step S101, need exist for carrying out human face region image to be processed comparatively accurate face lookup
Match, how to judge its fog-level, it is therefore desirable to select the higher face identification method of some accuracy to carry out recognition of face,
Including but not limited to such as some human face recognition models obtained based on deep learning model training, or it is based on human face characteristic point
The human face recognition model matched.The occurrence of face probability threshold value can be implemented by technical staff's sets itself, or with reference to the present invention
Example four is handled to obtain, and is not limited herein.
It acquires various types of fuzzy facial image sample in the embodiment of the present invention in advance first, and is based on these samples
The original training for carrying out Fuzzy Identification Model, to obtain to carry out the Fuzzy Identification Model of fuzzy image recognition, in this base
On plinth, on the one hand by carrying out facial contour cutting and picture size adjustment to image to be processed in advance, key person is both remained
Face feature, and the judgement obscured to facial image is enabled to concentrate on face itself, the clarity for reducing non-face region is dry
It disturbs, it is on the other hand, then by carrying out fuzzy diagnosis to human face region image to be processed based on trained Fuzzy Identification Model, real
Show and the compatibility of a variety of face vague category identifiers has been accurately identified, whether accurate judgement human face region to be processed obscures, in order to anti-
Only facial contour identification and interference that may be present in Fuzzy Identification Model identification cause identification error to increase, while meeting reality
The demand of application only picks out the higher facial image of fog-level, and the embodiment of the present invention is also further to being judged as
The image to be processed of blurred picture is made whether the secondary verification comprising face, and in comprising the lesser situation of face probability,
That is face fog-level it is more high-leveled and difficult normally to identify wherein face in the case where, just finally determine that image to be processed is fuzzy people
Face image, to greatly improve finally to the compatibility and accuracy rate of different types of fuzzy facial image identification.
As a kind of specific implementation for being trained building to Fuzzy Identification Model, as shown in Fig. 2, of the invention real
Before applying example one, the embodiment of the present invention two includes:
S201 obtains the fuzzy facial image sample of clear face image sample and multi-Fuzzy type.
Wherein to the acquisition modes of facial image sample, voluntarily clapped either being crawled from website and being also possible to technical staff
It takes the photograph, not limits herein, but it should guaranteeing includes light vague category identifier, motion blur in the fuzzy facial image sample obtained
The fuzzy facial image sample of type and resolution ratio vague category identifier, to guarantee the validity of the subsequent model trained, meanwhile, three
The quantity that seed type obscures facial image sample should be approximately or identical.
Obtain a kind of specific implementation of fuzzy facial image sample as the present invention, in the embodiment of the present invention three, by
Technical staff Image Acquisition and store to face under different scene conditions in advance, required different types of to obtain
Fuzzy facial image sample, as shown in Figure 3, comprising:
S301 obtains the first preset quantity light obtained under different light conditions to face progress Image Acquisition and obscures
The fuzzy facial image sample of type.
Wherein, in order to enrich the fuzzy facial image samples of different fog-levels, in the embodiment of the present invention it is adjustable not
Same light intensity changes the environment of acquisition, pedestrian's face image acquiring of going forward side by side using multi-light type simultaneously.
S302 obtains camera and is moving mould from the second preset quantity shot under a variety of different relative moving speeds of human body
Paste the fuzzy facial image sample of type.
Similarly, in the embodiment of the present invention, also adjustable different relative moving speed acquires facial image, to obtain
The corresponding different fog-levels of motion blur type obscure facial image sample.
S303 obtains clear face image sample, and carries out pixel Fuzzy processing to clear face image sample, obtains
The fuzzy facial image sample of third preset quantity resolution ratio vague category identifier.
Here pixel Fuzzy Processing mode includes but is not limited to as fuzzy using mean value, intermediate value is fuzzy, Gaussian Blur, double
Side mode paste and scaling are fuzzy to clear face image sample progress Fuzzy Processing, obtain the fuzzy of corresponding resolution ratio vague category identifier
Facial image sample.Wherein, in order to guarantee the balance between sample data, three types obscure face figure in the embodiment of the present invention
Decent quantity should be approximately or identical, specifically, the first preset quantity, the second preset quantity and third preset quantity
Value can be by technical staff's sets itself.
As a preferred embodiment of the present invention, on the basis of the embodiment of the present invention three, the embodiment of the present invention is also wrapped
It includes: crawling preset website, obtain the fuzzy facial image sample of the 4th preset quantity.Wherein, the 4th preset quantity, first pre-
If quantity, the second preset quantity and third preset quantity are sequentially reduced.
Since required sample size is more, whole shooting/processing difficulties are larger and sample randomness is inadequate, it is difficult to guarantee
Final effect, it is therefore desirable to crawl the fuzzy facial image being likely to occur in real life, but in actual conditions, can crawl
Fuzzy human face data inside include vague category identifier be uncontrollable, and according to the image data that can be crawled to actual website
It is found after being analyzed, various types of fuzzy facial image quantity distinguishes situation are as follows: resolution ratio vague category identifier > motion blur type
> light vague category identifier, therefore on the basis of actually crawling, in order to balance the fuzzy facial image number of three types as much as possible
Amount, in the embodiment of the present invention in demand to the first preset quantity, the second preset quantity and third preset quantity change,
Need to meet the 4th preset quantity > the first preset quantity > second preset quantity > third preset quantity, wherein specific quantity is big
It is small to be set by technical staff according to situation is actually crawled.
S202, respectively to clear face image sample and fuzzy facial image sample carry out facial contour cutting and adjust to
Preset image sizes obtain corresponding clear human face region image pattern and fuzzy human face region image pattern.
Due to needing to acquire a large amount of facial image sample in the embodiment of the present invention and carrying out model training, in order to reduce mould
The workload and speed of type training, in the embodiment of the present invention can to clear face image sample and fuzzy facial image sample standard deviation into
Row facial contour is cut, while need to carry out in embodiments of the present invention is fuzzy recognition of face rather than fine face knowledge
Other or feature extraction etc., therefore excessive human face region data are not needed, therefore the present invention is implemented when carrying out selection of dimension
Example preferably selects lesser size, on the basis of guaranteeing fuzzy diagnosis, increase as much as possible model training speed and
To the fuzzy facial image recognition speed of image to be processed in the embodiment of the present invention one.Wherein, specific facial contour cutting side
Method can refer to related description in the embodiment of the present invention one, and it will not go into details herein.
S203, based on clear human face region image pattern and fuzzy human face region image pattern to preset two disaggregated model
It is trained, and calculates the penalty values of classification results cross entropy.
As a kind of specific implementation for the penalty values for calculating classification results cross entropy in the embodiment of the present invention two, packet
It includes:
The penalty values of classification results cross entropy are calculated based on following formula (1):
Wherein, L is penalty values, and n is the total quantity of face area image sample, xiAnd yiIt is i-th human face region figure respectively
Decent physical tags value and tag along sort value, miFor the face quantity for including in i-th human face region image pattern, people
Face area image sample includes clear human face region image pattern and fuzzy human face region image pattern.
S204 is updated training to two disaggregated models based on gradient descent method if penalty values are greater than default loss threshold value,
Until penalty values are less than or equal to default loss threshold value or the number of iterations reaches preset times threshold value, training is completed, is obscured
Identification model.
Wherein, two disaggregated models specifically used not limit herein, including but not limited to such as logic-based regression model
Two disaggregated models and two disaggregated models based on machine learning etc., can be by technical staff's self-setting according to actual needs.
As the embodiment of the present invention four, it is contemplated that the embodiments of the present invention one to three are in the secondary inspection for carrying out blurred picture
When testing, the accuracy of inspection can be directly by the accuracy of used human face recognition model and used face probability threshold
It is worth availability influence, determines if directly setting a fixed face probability threshold value by technical staff, the accuracy of inspection
The subjectivity for being highly prone to technical staff influences, so that accuracy is difficult to be protected, therefore, in order to improve having for inspection
Effect property, as shown in Figure 4, comprising:
S401 carries out face knowledge to clear face image sample and fuzzy facial image sample based on human face recognition model
Not, every clear face image sample and the corresponding probability comprising face of fuzzy facial image sample are obtained.
S402, it is corresponding general comprising face based on every clear face image sample and fuzzy facial image sample
Rate determines default face probability threshold value.
In view of no matter which kind of face recognition algorithms is impossible to realize very accurate fuzzy diagnosis in practical application,
Effective and reasonable face probability threshold value in order to obtain simultaneously, can be practical selected using the embodiment of the present invention one in the embodiment of the present invention
Human face recognition model, to be identified to the facial image sample actually got, and by the result of identification and every practical
The label of facial image sample is compared, select can wherein make two classifying qualities optimal comprising the conduct of face probability
Face probability threshold value in the embodiment of the present invention one, for example, being obtained to the progress recognition of face of each facial image sample corresponding
After face probability, when choosing comprising face probability=70%, the effect of two classification is optimal, and the present invention is implemented at this time for discovery
Example will be by 70% as the face probability threshold value in the embodiment of the present invention one.
In the embodiment of the present invention, face probability threshold is carried out by the human face recognition model and facial image sample of actual use
The calculating of value is chosen, and can be avoided as much as different faces identification model identification accuracy height to identification caused by the application
The situation of inaccuracy, so that no matter technical staff chooses which type of face identification model, the embodiment of the present invention in practical application
Suitable threshold value can be determined to the recognition capability of actual sample according to human face recognition model well, so that this
The ability of a pair of clear recognition of face of inventive embodiments realizes adaptive optimization.
Corresponding to the method for foregoing embodiments, Fig. 5 shows fuzzy facial image identification dress provided in an embodiment of the present invention
The structural block diagram set, for ease of description, only parts related to embodiments of the present invention are shown.The exemplary fuzzy face of Fig. 5
Pattern recognition device can be the executing subject of the fuzzy facial image recognition method of the offer of previous embodiment one.
Referring to Fig. 5, which includes:
Face cuts module 51, for carrying out facial contour cutting to image to be processed and adjusting to preset image sizes,
Obtain corresponding human face region image to be processed.
Fuzzy image recognition module 52 is trained in advance fuzzy for the human face region image to be processed to be input to
It is handled in identification model, identifies whether the human face region image to be processed is blurred picture, wherein Fuzzy Identification Model
It is trained to be in advance based on the fuzzy human face region image pattern of clear human face region image pattern and multi-Fuzzy type
The identification model arrived, whether image is blurred picture for identification.
Face recognition module 53 is based on preset people if being blurred picture for the human face region image to be processed
Face identification model carries out recognition of face to the image to be processed, judges that the image to be processed includes the probability of face.
Fuzzy face recognition module 54, if the probability for the image to be processed to include face is less than default face probability
Threshold value then determines that the image to be processed is fuzzy facial image.
Further, the fuzzy facial image identification device, further includes:
Sample acquisition module, for obtaining the fuzzy facial image sample of clear face image sample and multi-Fuzzy type
This.
Sample cuts module, for carrying out respectively to the clear face image sample and the fuzzy facial image sample
Facial contour cuts and adjusts to the preset image sizes, obtains the corresponding clear human face region image pattern and described
Fuzzy human face region image pattern.
Model training module, for being based on the clear human face region image pattern and the fuzzy human face region image sample
This is trained preset two disaggregated model, and calculates the penalty values of classification results cross entropy.
Repetitive exercise module, if being greater than default loss threshold value for the penalty values, based on gradient descent method to described two
Disaggregated model is updated training, until the penalty values are less than or equal to the default loss threshold value or the number of iterations reaches pre-
If frequency threshold value, training is completed, the Fuzzy Identification Model is obtained.
Further, sample acquisition module, comprising:
It obtains and the first preset quantity light vague category identifier that Image Acquisition obtains is carried out to face under different light conditions
The fuzzy facial image sample.
Obtain camera from the second preset quantity motion blur class for being shot under a variety of different relative moving speeds of human body
The fuzzy facial image sample of type.
The clear face image sample is obtained, and pixel Fuzzy processing is carried out to the clear face image sample,
Obtain the fuzzy facial image sample of third preset quantity resolution ratio vague category identifier.
Further, sample acquisition module, further includes:
Preset website is crawled, the fuzzy facial image sample of the 4th preset quantity is obtained.Wherein, the 4th present count
Amount, the first preset quantity, the second preset quantity and third preset quantity are sequentially reduced.
Further, model training module, including
The penalty values of the classification results cross entropy are calculated based on following formula (1):
Wherein, L is penalty values, and n is the total quantity of face area image sample, xiAnd yiIt is i-th human face region figure respectively
Decent physical tags value and tag along sort value, miFor the face quantity for including in i-th human face region image pattern, people
Face area image sample includes clear human face region image pattern and fuzzy human face region image pattern.
Further, the fuzzy facial image identification device, further includes:
The clear face image sample and the fuzzy facial image sample are carried out based on the human face recognition model
Recognition of face, obtains every clear face image sample and the fuzzy facial image sample is corresponding comprising face
Probability.
Corresponding based on clear face image sample described in every and the fuzzy facial image sample includes face
Probability, determine the default face probability threshold value.
Each module realizes the process of respective function in fuzzy facial image identification device provided in an embodiment of the present invention, specifically
It can refer to the description of aforementioned embodiment illustrated in fig. 1 one, details are not described herein again.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Although will also be appreciated that term " first ", " second " etc. are used in some embodiment of the present invention in the text
Various elements are described, but these elements should not be limited by these terms.These terms are used only to an element
It is distinguished with another element.For example, the first table can be named as the second table, and similarly, the second table can be by
It is named as the first table, without departing from the range of various described embodiments.First table and the second table are all tables, but
It is them is not same table.
Fig. 6 is the schematic diagram for the terminal device that one embodiment of the invention provides.As shown in fig. 6, the terminal of the embodiment is set
Standby 6 include: processor 60, memory 61, and the computer that can be run on the processor 60 is stored in the memory 61
Program 62.The processor 60 realizes that above-mentioned each fuzzy facial image recognition method is implemented when executing the computer program 62
Step in example, such as step 101 shown in FIG. 1 is to 105.Alternatively, when the processor 60 executes the computer program 62
Realize the function of each module/unit in above-mentioned each Installation practice, such as the function of module 51 to 54 shown in Fig. 5.
The terminal device 6 can be the calculating such as desktop PC, notebook, palm PC and cloud server and set
It is standby.The terminal device may include, but be not limited only to, processor 60, memory 61.It will be understood by those skilled in the art that Fig. 6
The only example of terminal device 6 does not constitute the restriction to terminal device 6, may include than illustrating more or fewer portions
Part perhaps combines certain components or different components, such as the terminal device can also include input sending device, net
Network access device, bus etc..
Alleged processor 60 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng.
The memory 61 can be the internal storage unit of the terminal device 6, such as the hard disk or interior of terminal device 6
It deposits.The memory 61 is also possible to the External memory equipment of the terminal device 6, such as be equipped on the terminal device 6
Plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card dodge
Deposit card (Flash Card) etc..Further, the memory 61 can also both include the storage inside list of the terminal device 6
Member also includes External memory equipment.The memory 61 is for storing needed for the computer program and the terminal device
Other programs and data.The memory 61, which can be also used for temporarily storing, have been sent or data to be sent.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated module/unit be realized in the form of SFU software functional unit and as independent product sale or
In use, can store in a computer readable storage medium.Based on this understanding, the present invention realizes above-mentioned implementation
All or part of the process in example method, can also instruct relevant hardware to complete, the meter by computer program
Calculation machine program can be stored in a computer readable storage medium, the computer program when being executed by processor, it can be achieved that on
The step of stating each embodiment of the method.Wherein, the computer program includes computer program code, the computer program generation
Code can be source code form, object identification code form, executable file or certain intermediate forms etc..The computer-readable medium
It may include: any entity or device, recording medium, USB flash disk, mobile hard disk, magnetic that can carry the computer program code
Dish, CD, computer storage, read-only memory (Read-Only Memory, ROM), random access memory (Random
Access Memory, RAM), electric carrier signal, telecommunication signal and software distribution medium etc..
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features；And these are modified
Or replacement, the essence of corresponding technical solution is departed from the spirit and scope of the technical scheme of various embodiments of the present invention, it should all
It is included within protection scope of the present invention.
Claims (10)
1. a kind of fuzzy facial image recognition method characterized by comprising
Facial contour cutting is carried out to image to be processed and is adjusted to preset image sizes, corresponding human face region to be processed is obtained
Image；
The human face region image to be processed is input in preparatory trained Fuzzy Identification Model and is handled, described in identification
Whether human face region image to be processed is blurred picture, wherein Fuzzy Identification Model is to be in advance based on clear human face region image
The identification model that sample and the fuzzy human face region image pattern of multi-Fuzzy type are trained, for identification image be
No is blurred picture；
If the human face region image to be processed is blurred picture, based on preset human face recognition model to the figure to be processed
As carrying out recognition of face, judge that the image to be processed includes the probability of face；
If the probability that the image to be processed includes face is less than default face probability threshold value, determine that the image to be processed is
Fuzzy facial image.
2. fuzzy facial image recognition method as described in claim 1, which is characterized in that the instruction of the Fuzzy Identification Model
White silk includes:
Obtain the fuzzy facial image sample of clear face image sample and multi-Fuzzy type；
Respectively to the clear face image sample and the fuzzy facial image sample carry out facial contour cutting and adjust to
The preset image sizes obtain the corresponding clear human face region image pattern and the fuzzy human face region image sample
This；
Based on the clear human face region image pattern and the fuzzy human face region image pattern to preset two disaggregated model
It is trained, and calculates the penalty values of classification results cross entropy；
If the penalty values are greater than default loss threshold value, training is updated to two disaggregated model based on gradient descent method,
Until the penalty values are less than or equal to the default loss threshold value or the number of iterations reaches preset times threshold value, training is completed,
Obtain the Fuzzy Identification Model.
3. fuzzy facial image recognition method as claimed in claim 2, which is characterized in that the vague category identifier includes movement mould
Paste, light are fuzzy and resolution ratio is fuzzy, the fuzzy face figure for obtaining clear face image sample and multi-Fuzzy type
Decent, comprising:
Obtain the institute for carrying out the first preset quantity light vague category identifier that Image Acquisition obtains to face under different light conditions
State fuzzy facial image sample；
Obtain camera from the second preset quantity motion blur type for being shot under a variety of different relative moving speeds of human body
The fuzzy facial image sample；
The clear face image sample is obtained, and pixel Fuzzy processing is carried out to the clear face image sample, is obtained
The fuzzy facial image sample of third preset quantity resolution ratio vague category identifier.
4. fuzzy facial image recognition method as claimed in claim 3, which is characterized in that the vague category identifier includes movement mould
Paste, light are fuzzy and resolution ratio is fuzzy, the fuzzy face figure for obtaining clear face image sample and multi-Fuzzy type
Decent, further includes:
Preset website is crawled, the fuzzy facial image sample of the 4th preset quantity is obtained；Wherein, the 4th preset quantity,
First preset quantity, the second preset quantity and third preset quantity are sequentially reduced.
5. fuzzy facial image recognition method as claimed in claim 2, which is characterized in that the calculating classification results cross entropy
Penalty values, including
The penalty values of the classification results cross entropy are calculated based on following formula:
Wherein, L is penalty values, and n is the total quantity of face area image sample, xiAnd yiIt is i-th human face region image sample respectively
This physical tags value and tag along sort value, miFor the face quantity for including in i-th human face region image pattern, face area
Area image sample includes clear human face region image pattern and fuzzy human face region image pattern.
6. the fuzzy facial image recognition method as described in claim 2 to 5 any one, which is characterized in to be processed
Image carries out facial contour cutting and adjusts to before preset image sizes, further includes:
Face is carried out to the clear face image sample and the fuzzy facial image sample based on the human face recognition model
Identification, obtains every clear face image sample and the fuzzy facial image sample is corresponding general comprising face
Rate；
It is corresponding general comprising face based on clear face image sample described in every and the fuzzy facial image sample
Rate determines the default face probability threshold value.
7. a kind of fuzzy facial image identification device characterized by comprising
Face cuts module, for facial contour cutting and adjust to preset image sizes to image to be processed, obtains pair
The human face region image to be processed answered；
Fuzzy image recognition module, for the human face region image to be processed to be input to preparatory trained fuzzy diagnosis mould
It is handled in type, identifies whether the human face region image to be processed is blurred picture, wherein Fuzzy Identification Model is preparatory
The knowledge that fuzzy human face region image pattern based on clear human face region image pattern and multi-Fuzzy type is trained
Other model, whether image is blurred picture for identification；
Face recognition module is based on preset recognition of face if being blurred picture for the human face region image to be processed
Model carries out recognition of face to the image to be processed, judges that the image to be processed includes the probability of face；
Fuzzy face recognition module, if the probability for the image to be processed to include face is less than default face probability threshold value,
Then determine that the image to be processed is fuzzy facial image.
8. fuzzy facial image identification device as claimed in claim 7, which is characterized in that further include:
Sample acquisition module, for obtaining the fuzzy facial image sample of clear face image sample and multi-Fuzzy type；
Sample cuts module, for carrying out face to the clear face image sample and the fuzzy facial image sample respectively
Profile cuts and adjusts to the preset image sizes, obtains the corresponding clear human face region image pattern and described fuzzy
Human face region image pattern；
Model training module, for being based on the clear human face region image pattern and the fuzzy human face region image pattern pair
Preset two disaggregated model is trained, and calculates the penalty values of classification results cross entropy；
Repetitive exercise module, if being greater than default loss threshold value for the penalty values, based on gradient descent method to two classification
Model is updated training, until the penalty values are less than or equal to the default loss threshold value or the number of iterations reaches default time
Number threshold value, completes training, obtains the Fuzzy Identification Model.
9. a kind of terminal device, which is characterized in that the terminal device includes memory, processor, is stored on the memory
There is the computer program that can be run on the processor, is realized when the processor executes the computer program as right is wanted
The step of seeking any one of 1 to 6 the method.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In when the computer program is executed by processor the step of any one of such as claim 1 to 6 of realization the method.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910536758.XA CN110378235B (en) | 2019-06-20 | 2019-06-20 | Fuzzy face image recognition method and device and terminal equipment |
| PCT/CN2019/103159 WO2020252917A1 (en) | 2019-06-20 | 2019-08-29 | Fuzzy face image recognition method and apparatus, terminal device, and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910536758.XA CN110378235B (en) | 2019-06-20 | 2019-06-20 | Fuzzy face image recognition method and device and terminal equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110378235A true CN110378235A (en) | 2019-10-25 |
| CN110378235B CN110378235B (en) | 2024-05-28 |
Family
ID=68250416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910536758.XA Active CN110378235B (en) | 2019-06-20 | 2019-06-20 | Fuzzy face image recognition method and device and terminal equipment |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110378235B (en) |
| WO (1) | WO2020252917A1 (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110992395A (en) * | 2019-11-01 | 2020-04-10 | 北京达佳互联信息技术有限公司 | Image training sample generation method and device and motion tracking method and device |
| CN111222411A (en) * | 2019-11-28 | 2020-06-02 | 中国船舶重工集团公司第七一三研究所 | Laser emission safe and rapid alarm method and device |
| CN111460939A (en) * | 2020-03-20 | 2020-07-28 | 深圳市优必选科技股份有限公司 | Deblurring face recognition method and system and inspection robot |
| CN111814702A (en) * | 2020-07-13 | 2020-10-23 | 安徽兰臣信息科技有限公司 | Child face recognition method based on adult face and child photo feature space mapping relation |
| CN112001940A (en) * | 2020-08-21 | 2020-11-27 | Oppo（重庆）智能科技有限公司 | Image processing method and device, terminal and readable storage medium |
| CN112395965A (en) * | 2020-11-06 | 2021-02-23 | 山东鲁能软件技术有限公司 | Mobile terminal face recognition system and method based on power intranet |
| CN112669207A (en) * | 2020-12-21 | 2021-04-16 | 四川长虹电器股份有限公司 | Method for enhancing resolution of face image based on television camera |
| CN112965604A (en) * | 2021-03-29 | 2021-06-15 | 深圳市优必选科技股份有限公司 | Gesture recognition method and device, terminal equipment and computer readable storage medium |
| WO2021189827A1 (en) * | 2020-08-06 | 2021-09-30 | 平安科技（深圳）有限公司 | Method and apparatus for recognizing blurred image, and device and computer-readable storage medium |
| CN113705461A (en) * | 2021-08-30 | 2021-11-26 | 平安银行股份有限公司 | Face definition detection method, device, equipment and storage medium |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112949423B (en) * | 2021-02-07 | 2024-05-24 | 深圳市优必选科技股份有限公司 | Object recognition method, object recognition device and robot |
| CN112766235B (en) * | 2021-02-24 | 2024-07-05 | 北京嘀嘀无限科技发展有限公司 | Face recognition method, apparatus, device, storage medium and computer program product |
| CN113436137A (en) * | 2021-03-12 | 2021-09-24 | 北京世纪好未来教育科技有限公司 | Image definition recognition method, device, equipment and medium |
| CN113657178B (en) * | 2021-07-22 | 2024-08-06 | 浙江大华技术股份有限公司 | Face recognition method, electronic device and computer readable storage medium |
| CN113706402A (en) * | 2021-07-23 | 2021-11-26 | 维沃移动通信（杭州）有限公司 | Neural network training method and device and electronic equipment |
| CN114549346A (en) * | 2022-01-27 | 2022-05-27 | 阿丘机器人科技(苏州)有限公司 | Blurred image recognition method, device, equipment and storage medium |
| CN114648672A (en) * | 2022-02-25 | 2022-06-21 | 北京百度网讯科技有限公司 | Method and device for constructing sample image set, electronic equipment and readable storage medium |
| CN115171197B (en) * | 2022-09-01 | 2023-05-16 | 广州市森锐科技股份有限公司 | High-precision image information identification method, system, equipment and storage medium |
| CN117240607B (en) * | 2023-11-10 | 2024-02-13 | 北京云尚汇信息技术有限责任公司 | Security authentication method based on security computer |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108241821A (en) * | 2016-12-23 | 2018-07-03 | 北京三星通信技术研究有限公司 | Image processing equipment and method |
| CN109829396A (en) * | 2019-01-16 | 2019-05-31 | 广州杰赛科技股份有限公司 | Recognition of face motion blur processing method, device, equipment and storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10726244B2 (en) * | 2016-12-07 | 2020-07-28 | Samsung Electronics Co., Ltd. | Method and apparatus detecting a target |
| CN106951825B (en) * | 2017-02-13 | 2021-06-29 | 苏州飞搜科技有限公司 | Face image quality evaluation system and implementation method |
| CN107133948B (en) * | 2017-05-09 | 2020-05-08 | 电子科技大学 | Image blurring and noise evaluation method based on multitask convolution neural network |
| CN108288027B (en) * | 2017-12-28 | 2020-10-27 | 新智数字科技有限公司 | Image quality detection method, device and equipment |
-
2019
- 2019-06-20 CN CN201910536758.XA patent/CN110378235B/en active Active
- 2019-08-29 WO PCT/CN2019/103159 patent/WO2020252917A1/en active Application Filing
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108241821A (en) * | 2016-12-23 | 2018-07-03 | 北京三星通信技术研究有限公司 | Image processing equipment and method |
| CN109829396A (en) * | 2019-01-16 | 2019-05-31 | 广州杰赛科技股份有限公司 | Recognition of face motion blur processing method, device, equipment and storage medium |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110992395A (en) * | 2019-11-01 | 2020-04-10 | 北京达佳互联信息技术有限公司 | Image training sample generation method and device and motion tracking method and device |
| CN110992395B (en) * | 2019-11-01 | 2023-08-18 | 北京达佳互联信息技术有限公司 | Image training sample generation method and device and motion tracking method and device |
| CN111222411A (en) * | 2019-11-28 | 2020-06-02 | 中国船舶重工集团公司第七一三研究所 | Laser emission safe and rapid alarm method and device |
| CN111222411B (en) * | 2019-11-28 | 2023-09-01 | 中国船舶重工集团公司第七一三研究所 | Laser emission safety rapid alarm method and device |
| CN111460939A (en) * | 2020-03-20 | 2020-07-28 | 深圳市优必选科技股份有限公司 | Deblurring face recognition method and system and inspection robot |
| CN111814702A (en) * | 2020-07-13 | 2020-10-23 | 安徽兰臣信息科技有限公司 | Child face recognition method based on adult face and child photo feature space mapping relation |
| WO2021189827A1 (en) * | 2020-08-06 | 2021-09-30 | 平安科技（深圳）有限公司 | Method and apparatus for recognizing blurred image, and device and computer-readable storage medium |
| CN112001940A (en) * | 2020-08-21 | 2020-11-27 | Oppo（重庆）智能科技有限公司 | Image processing method and device, terminal and readable storage medium |
| CN112395965A (en) * | 2020-11-06 | 2021-02-23 | 山东鲁能软件技术有限公司 | Mobile terminal face recognition system and method based on power intranet |
| CN112669207A (en) * | 2020-12-21 | 2021-04-16 | 四川长虹电器股份有限公司 | Method for enhancing resolution of face image based on television camera |
| CN112965604A (en) * | 2021-03-29 | 2021-06-15 | 深圳市优必选科技股份有限公司 | Gesture recognition method and device, terminal equipment and computer readable storage medium |
| CN113705461A (en) * | 2021-08-30 | 2021-11-26 | 平安银行股份有限公司 | Face definition detection method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110378235B (en) | 2024-05-28 |
| WO2020252917A1 (en) | 2020-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110378235A (en) | A kind of fuzzy facial image recognition method, device and terminal device | |
| CN108197532B (en) | The method, apparatus and computer installation of recognition of face | |
| KR102319177B1 (en) | Method and apparatus, equipment, and storage medium for determining object pose in an image | |
| US10936911B2 (en) | Logo detection | |
| CN103927387B (en) | Image indexing system and its correlation technique and device | |
| US9396560B2 (en) | Image-based color palette generation | |
| CN109284733B (en) | Shopping guide negative behavior monitoring method based on yolo and multitask convolutional neural network | |
| US8571332B2 (en) | Methods, systems, and media for automatically classifying face images | |
| CN110379020B (en) | Laser point cloud coloring method and device based on generation countermeasure network | |
| CN110070101A (en) | Floristic recognition methods and device, storage medium, computer equipment | |
| CN110222686B (en) | Object detection method, object detection device, computer equipment and storage medium | |
| CN108416902A (en) | Real-time object identification method based on difference identification and device | |
| CN111008935B (en) | Face image enhancement method, device, system and storage medium | |
| CN113705461B (en) | Face definition detection method, device, equipment and storage medium | |
| CN110135889A (en) | Method, server and the storage medium of intelligent recommendation book list | |
| CN110827312A (en) | Learning method based on cooperative visual attention neural network | |
| WO2020230244A1 (en) | Training method, training program, and training device | |
| CN109977875A (en) | Gesture identification method and equipment based on deep learning | |
| CN108427971A (en) | The method and system of tobacco leaf grading based on mobile terminal | |
| CN111382791B (en) | Deep learning task processing method, image recognition task processing method and device | |
| CN109726756A (en) | Image processing method, device, electronic equipment and storage medium | |
| CN109815823A (en) | Data processing method and Related product | |
| CN114494751A (en) | License information identification method, device, equipment and medium | |
| CN109583492A (en) | A kind of method and terminal identifying antagonism image | |
| CN115471439A (en) | Method and device for identifying defects of display panel, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |