CN108256459A - Library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically - Google Patents
Library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically Download PDFInfo
- Publication number
- CN108256459A CN108256459A CN201810021107.2A CN201810021107A CN108256459A CN 108256459 A CN108256459 A CN 108256459A CN 201810021107 A CN201810021107 A CN 201810021107A CN 108256459 A CN108256459 A CN 108256459A
- Authority
- CN
- China
- Prior art keywords
- face
- detector gate
- algorithm
- database
- automatically
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 14
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims abstract description 12
- 238000001303 quality assessment method Methods 0.000 claims abstract 3
- 238000012216 screening Methods 0.000 claims abstract 2
- 210000000887 face Anatomy 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 claims 1
- 238000010606 normalization Methods 0.000 claims 1
- 230000006870 function Effects 0.000 description 10
- 239000011159 matrix material Substances 0.000 description 5
- 239000002184 metal Substances 0.000 description 5
- 230000001815 facial effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000034 method Methods 0.000 description 3
- 230000036544 posture Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013441 quality evaluation Methods 0.000 description 2
- 241001533251 Gutierrezia sarothrae Species 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000003475 lamination Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 210000000162 simple eye Anatomy 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
- Image Processing (AREA)
Abstract
The present invention is in order to solve the problems, such as detector gate quick security check, it is proposed that library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically.What the algorithm of the present invention did not needed to tested personnel cooperates with the brush face of formula on one's own initiative, and only needs to walk according to normal route, belongs to non-formula face recognition algorithms.Multiple video cameras are installed along the direction that people enters on detector gate, these video cameras acquire video and carry out Face datection simultaneously, the face detected is screened by the laggard pedestrian's face quality assessment modules of face tracking, face after screening selects the face for belonging to same person by corresponding upper part of the body matching algorithm, the excessive face of angle is deleted by Attitude estimation again, depth convolutional neural networks extraction feature will be then inputted after face alignment.Based on the face characteristic extracted above, face matching is carried out by multiple-camera face alignment algorithm and builds library automatically.The detector gate face recognition algorithms of the multiple-camera fusion of the present invention, have been obviously improved the speed of safety check, alleviate due to receiving the security risks such as crowded caused by safety check.
Description
Technical field
The invention belongs to safety-security areas and field of safety check, are related to pattern-recognition, graph and image processing, machine learning etc., profit
With multi-camera acquisition face, and by these faces by human face modeling and based on the face recognition technology of deep learning into
Row fusion, then establishes multiple-camera recognition of face library automatically.
Background technology
The world seems peace, but and it is dangerous, terrorism and extremism quiet penetrate into the life of people
In work, so safety precaution is still a very important subject.Safety check can detect dangerous material in time to a certain extent, protect
Hinder the security of the lives and property of numerous people, such as airport, subway, important venue etc. have deployed safety check and set in various public arenas
It is standby.Detector gate is a kind of equipment being detected to human body, and the country mainly takes human body safety check traveling type metal detection door, auxiliary
Suspicious object is found in the form of the personal scanning of portable hand-held metal detector and safety inspector by hand " clap, touch, by, pressure ", such as
Cutter etc..This detector gate only has metallics reaction, and helpless for other contrabands, even to metal
The detection of substance, positioning is also very coarse, so the human body with big part metal can only be detected.And mating hand-held metal
Survey meter needs security staff to contact by security staff, this is easy to cause by the discontented mood of security staff or even causes limb
Body conflict, and this hand is examined, and generally requires 6~8 seconds/people, this safety check speed many occasions easily cause congestion and
It is detained, such as subway.
There are many producers to introduce recognition of face in safe examination system, then generally use brush identity card allows tested personnel to match
Brush face is closed, the face that the face of identity card and tested personnel brush out is compared, i.e. testimony of a witness verifying system.Testimony of a witness verifying system
It is the system that a kind of user cooperates with formula on one's own initiative, user is needed to acquire positive face against face collecting device (video camera).In addition, it is desirable to
User carries identity card, this may not necessarily many times accomplish, particularly hastily for as the passenger in subway system
Even if the working clan to hurry on a journey with identity card, needs to take out identity card in packet, then cooperates with face acquisition, this mistake on one's own initiative
The time of journey consumption is also long, and this safety check speed can not also meet as the great situation of this volume of the flow of passengers of subway.
Invention content
The present invention is in order to solve the problems, such as detector gate quick security check, it is proposed that the detector gate face based on multiple-camera fusion
Identification and face build library algorithm automatically.What the algorithm of the present invention did not needed to tested personnel cooperates with the brush face of formula on one's own initiative, and only needs
It walks according to normal route, belongs to non-formula face recognition algorithms.The face recognition algorithms of current mainstream on the market are to people
The posture of face all requires, and posture is more positive, then the accuracy rate of recognizer is higher.Using multiple and different position different angles
Camera shooting function captures the face of different heights and different postures of walking, it is ensured that higher positive face candid photograph rate, so the present invention exists
N number of video camera (N >=3) is installed on detector gate.N number of video camera collected video is carried out simultaneously Face datection and face with
Track in the face queue obtained from every road video camera face tracking, (is passed by the flat of detector gate within the time cycle of restriction by people
The equal time is calculated), Automatic sieve selects face quality score highest face M (M≤N), and 1 people is at most filtered out per road
Face.Corresponding upper part of the body image is intercepted by face, the matching of edge and color is carried out to upper part of the body image, successful match is then recognized
To be the face collected from same person, using human face modeling algorithm calculate these faces horizontally rotate angle,
Pitch angle, inclination angle filter out face K (K≤M) of facial angle smaller (close to positive face).This K faces are passed through into spy
After sign point carries out alignment, input to depth convolutional neural networks and carry out face characteristic extraction, every face corresponds to one
1024 dimensional feature vectors, the corresponding feature vector of face common K of extraction K.By K feature vector respectively with the people in face database
Face feature vector is compared, if K face matching values in face database have more than or equal to first threshold, selects the first
Match and the highest people of matching value is as the final output identified, while the first matched this is captured face and is added to correspondence
Face database in, update face database.If K face matching values in face database are both less than first threshold, the is set up
Two threshold values, more than or equal to second threshold, then it is assumed that temporarily matching, every face that K is opened in faces are selected from face database successively
The most people of matching number (everyone corresponds to multiple storage faces), matching number is L respectively1,...,LK, corresponding average
It is S respectively with value1,...,SK, comprehensive score Score is matched by calculating, sorts according to Score, selects highest Score corresponding
People matches output, while this corresponding face captured is added in corresponding face database as final face, updates
Face database.If K matching values of the face in face database are both less than second threshold, new staff list is established,
This K faces add the demographic data as newly-built face of the personnel in face database in face database.The present invention passes through
The face recognition algorithms and build library algorithm automatically that multiple-camera merges make detector gate recognition of face be captured to the greatest extent to just
Face, and without being coordinated by security staff's brush face, the speed by detector gate is substantially increased, while according to recognition of face knot
Fruit establishes the face database by detector gate personnel automatically, and effective face database is provided for follow-up face management.
Library algorithm is built in detector gate recognition of face and face provided by the invention based on multiple-camera fusion automatically, including:
Initialize face database, can allow face database be it is empty or by the use of the face of collected by hand as initial bottom library,
The database indexes such as face picture, face characteristic, personal information are established, and interrelated.Each detector gate is according to storage ten million
Grade face quantity carries out the predistribution of disk and the predistribution of database.
N number of video camera, the direction that camera lens enters towards people are installed in both sides and top on detector gate, and people passes through from detector gate
When, positive face can be captured as far as possible.Fig. 1 is the detector gate schematic diagram for being mounted with 5 video cameras, and two have respectively been filled in the both sides of door
A video camera, top have filled a video camera.N number of video camera opens Face datection algorithm simultaneously.The present invention is not using based on deep
The Face datection algorithm of study is spent, primarily directed to the scene of detector gate, the volume of the flow of passengers is very big, it is desirable that there is very high detection speed,
And the Face datection based on deep learning, can not meet the needs of N number of video camera is carried out at the same time Face datection.In addition, in safety check
Door multiple-camera scene, positive face occurrence rate is very high, and background is relatively uniform, and is not belonging to completely without the people under constraints
Face detects, and therefore, the present invention uses follow-on Adaboost algorithm based on Haar-like features, both ensure that detection speed
Degree, in turn ensures very high positive face verification and measurement ratio and very low false drop rate.Since traditional Haar-like features are all neighborhood parts
Feature, such as simple eye feature formed with periphery, the present invention is in order to supplement the feature of greater room range, it is proposed that 3x3 structures
Haar-like features can preferably express the feature that face eyes, nose and mouth etc. are combined.Negative sample is adopted during training
Article such as luggage and various texture clothes for being likely to occur in the scene picture combination scene shot with detector gate environment etc..
After N number of video camera detects face, to the face of the appearance of each video camera into line trace, track algorithm uses
Kalman filter tracking algorithm based on neighborhood search.By face size constraint, the face of distant place is shielded, passes through detector gate
Face will not usually block, and typically be passed through one by one by detector gate, in such a scenario, it is assumed that face moves
Process noise and observation noise be all white Gaussian noise, this assume substantially set up, therefore Kalman filter tracking calculate
Method is effective under this scene.
The result of face tracking is stored in face queue, and detector gate is to ask for help to pass through one by one under normal conditions, therefore
Effective face queue of each video camera there is usually one.It (is passed by being averaged of detector gate by people within the time cycle of restriction
Time is calculated), face quality score highest face M (M≤N) is filtered out from the face queue of N number of video camera, per road
At most filter out 1 face.Face quality evaluation is carried out by the overall target to illuminance and clarity.Illuminance passes through face
The combined indexs such as the global average brightness in portion, maximum brightness value, brightness minimum, using the Y-component of yuv space.Clarity is led to
High fdrequency component is crossed to assess, first passes around discrete cosine transform (DCT), is then estimated by counting the number accounting of high frequency coefficient
Count clarity.
M derived above faces intercept corresponding upper part of the body image, to upper part of the body image calculating office by face coordinate
Portion's binary pattern (LBP) feature, while statistical color histogram.What LBP features and color histogram were expressed herein is that people wears
The features such as clothes and hair, are matched two-by-two by LBP features and color histogram, and finishing screen is selected to be belonged in M faces
The face of same person.The face filtered out is transmitted to human face modeling algorithm:Human face characteristic point extraction, feature are carried out first
Point includes the positions such as eyes, nose, the corners of the mouth, chin, this characteristic point is transmitted to the face alignment of next step simultaneously;Pass through face
For three-dimensional rotation model to the projection (matrix of three angles) of two dimension, estimate face horizontally rotates angle, pitch angle, inclination
Three, angle etc. angle, the face that angle is less than to certain threshold value (close to positive face) screen K (K≤M).
After this K alignments of the face by feature based point, depth convolutional neural networks extraction feature is inputed to,
Characteristic dimension is 1024 dimensions.As shown in figure 4, depth convolutional neural networks are by 9 convolutional layers, 4 pond layers, 1 merging layer and 1
A full articulamentum composition.Convolutional layer uses the convolution kernel of 3x3, and pond layer uses the window of 2x2, merges layer by different convolutional layers
Fusion Features.Closely follow ReLU (Rectified Linear Units) unit in each convolutional layer back.Every layer is returned feature
One changes.The function that final valuation functions are weighted using Softmax loss functions and center loss function.It will online disclosed star
Face database and the collected face of detector gate are trained by calibrated database, ultimately generate extraction face characteristic
Convolutional neural networks parameter.
K feature vector is compared respectively with the face feature vector in face database, if K faces are in face database
Middle matching value has more than or equal to first threshold, then selects the first matching and the highest people of matching value is as the final defeated of identification
Go out, while the first matched this is captured face and is added in corresponding face database, update face database.
If K face matching values in face database are both less than first threshold, second threshold is set up, more than or equal to second
Threshold value, then it is assumed that temporarily matching, it is (every that every face that K is opened in faces selects the most people of matching number from face database successively
Multiple corresponding storage faces of individual), matching number is L respectively1,...,LK, corresponding Average match is S respectively1,...,SK,
Comprehensive score is matched by calculating
It sorts according to Score, the corresponding people of highest Score is selected to match output as final face, while corresponding
The face of this candid photograph is added in corresponding face database, updates face database.
If K matching values of the face in face database are both less than second threshold, new staff list is established, this K
Face is opened as newly-built face of the personnel in face database, and the demographic data is added in face database.
Library algorithm is built in detector gate recognition of face and face proposed by the present invention based on multiple-camera fusion automatically, by more
A video camera multi-angle acquires face simultaneously, in the case that completely without cooperating on one's own initiative by security staff, it is ensured that face
The positive face rate captured, improves the safety check speed of detector gate, while improve the discrimination of the non-formula recognition of face of detector gate.
It can establish automatically by the face database of security staff again, the management for follow-up personnel provides face database support.
Description of the drawings
Fig. 1 is detector gate multiple-camera scheme of installation in the present invention.
Fig. 2 is that the detector gate recognition of face merged the present invention is based on multiple-camera and face build library algorithm flow chart automatically.
Fig. 3 is the increased Haar-like features schematic diagram of the present invention.
Fig. 4 is that each layer of depth convolutional neural networks of face feature extraction of the present invention illustrates schematic diagram.
Specific embodiment
The present invention is further expalined with specific example below in conjunction with the accompanying drawings.It is it should be noted that described below
Example be intended to be better understood from the present invention, only the present invention in a part, protection model not thereby limiting the invention
It encloses.
As shown in Fig. 2, the present invention realizes that being carried out at the same time Face datection, face tracking, face quality by multiple video cameras comments
Estimate, the extraction of human face modeling, face alignment, face characteristic, face characteristic compares, face builds a series of steps such as library automatically
Suddenly.
In step 201, create empty face database or by the use of the face of collected by hand as initial bottom library, establish
Face relevant information data library indexes.Disk storage space and memory headroom are distributed for each detector gate face identification system.
In step 202, N number of video camera is installed at both sides and top on detector gate, the direction that camera lens enters towards people, and people is from peace
Examine door by when, positive face can be captured as far as possible.Camera lens uses large aperture camera lens, reduces the time for exposure, to capture in movement
Face.N number of video camera acquires video and carries out Face datection simultaneously, and the present invention is using follow-on special based on Haar-like
The Adaboost algorithm of sign that is, using the feature of greater room range, increases the Haar-like features of 3x3 structures, can be more preferable
The feature combined such as expression face eyes, nose and mouth, as shown in Figure 3.Training when negative sample from subway detector gate,
The scenes such as airport security door, important venue detector gate are chosen.
Step 203, after step 202 detects face, to the face of the appearance of each video camera into line trace, tracking is calculated
Method uses the Kalman filter tracking algorithm based on neighborhood search.Using target's center as neighborhood search starting point, pass through position and speed
Degree prediction, the face detected in the nearest step 202 of detection range current face in a certain range of window, to search
Face coordinate for observation coordinate, pass through Kalman filter and update the variables such as position and speed.If not searching face,
Then it is considered the face to disappear, updates deletion frame by frame from face queue.If emerging face, then establish new face with
Track queue.
Step 204 (is passed by the result face queue of step 203 face tracking within the time cycle of restriction by people
The average time of detector gate is calculated), filter out the highest face M of face quality score from the face queue of N number of video camera
It opens (M≤N), 1 face is at most filtered out per road.Face quality evaluation by the overall target to illuminance and clarity into
Row, if the global average brightness I of faceAVG, maximum brightness value IMAX, brightness minimum IMIN, face global illumination degree index is such as
Under:
In brightness, to consider the influence of negative and positive face, the present invention is by the average brightness of left face and right face by following public
Formula is estimated:
The number accounting of high frequency coefficient after face clarity discrete cosine transform (DCT) counts:First image is divided
For the macro block of 8x8, the frequency domain Matrix C of DCT generations 8x8 is carried out to each macro block, to position c each in frequency domain matrixij(direct current point
Except amount, 1≤i, j≤8) all set up a threshold value TijIf the frequency domain matrix current location c after DCTijMore than Tij, then high frequency
The counter of component increases by one, and the high fdrequency component total number for eventually exceeding threshold value accounts for the ratio of frequency coefficient total number as face
Articulation index QUALITYsharpness.Therefore, face quality comprehensive index is as follows:
QUALITYface=(QUALITYbrightness+QUALITYuniformity+QUALITYsharpness)/3 × 100%
On the M that step 205 and step 206 are obtained in step 204 faces, the corresponding upper part of the body is intercepted by face coordinate
Image calculates the upper part of the body image local binary patterns (LBP) feature, while statistical color histogram.LBP features and color are straight
What side's figure was expressed herein is the features such as clothes and hair that people wears, is matched two-by-two by LBP features and color histogram, most
The face for belonging to same person in M faces is filtered out eventually.The face filtered out is transmitted to human face modeling algorithm:First into
Pedestrian's face characteristic point extracts, and characteristic point includes the positions such as eyes, nose, the corners of the mouth, chin.Human face characteristic point extraction is using convolution god
Through network:5 convolutional layers, 3 pond layers and 1 full articulamentum.This characteristic point is transmitted to the face alignment of next step simultaneously;
By face three-dimensional rotation model to the projection (matrix of three angles) of two dimension, horizontally rotating angle, bowing for face is estimated
Three angles such as the elevation angle, inclination angle, the face that angle is less than to certain threshold value (close to positive face) screen K (K≤M).
By step 206, this K opens alignments of the face by feature based point to step 207, and feature point coordinates is by step 206
Facial feature points detection module provides.Point, holding on the basis of the position coordinates such as eyes, nose, the corners of the mouth, chin bottom during calibration
The relative position of these datum marks is constant, and facial image is cut and zooms to fixed resolution, the present invention using
The face resolution ratio of 128x112.
Step 208, the facial image after alignment is inputed into depth convolutional neural networks extraction feature, characteristic dimension
For 1024 dimensions.Depth convolutional neural networks merge layer by 9 convolutional layers, 4 pond layers, 1 and 1 full articulamentum forms.Volume
Lamination uses the convolution kernel of 3x3, and pond layer uses the window of 2x2, merging layer by the 11st and the 12nd layer of Fusion Features, export to
Next layer.Closely follow ReLU (Rectified Linear Units) unit in each convolutional layer back.Feature is carried out normalizing by every layer
Change.The function that final valuation functions are weighted using Softmax loss functions and center loss function, center loss function choose compared with
Small weight.Online disclosed star's face database and the collected face of detector gate are carried out by calibrated database
Training, training are carried out by the way of interactive iteration, ultimately generate the convolutional neural networks parameter of extraction face characteristic.Training sample
This is made of 41000 people, altogether 500,000 faces.
Step 209 and step 210 K feature vector are compared respectively with the face feature vector in face database, such as
K face matching values in face database of fruit have more than or equal to first threshold, then select the first matching and the highest people of matching value
As the final output of identification, while the first matched this is captured face and is added in corresponding face database, update face
Database.
Step 211 and step 212 if K face matching values in face database are both less than first threshold, set up second
A threshold value, more than or equal to second threshold, then it is assumed that temporarily matching, every face in K faces are selected from face database successively
With the most people of number (everyone corresponds to multiple storage faces), matching number is L respectively1,...,LK, corresponding Mean match
Value is S respectively1,...,SK, comprehensive score is matched by calculating:
It sorts according to Score, the corresponding people of highest Score is selected to match output as final face, while corresponding
The face of this candid photograph is added in corresponding face database, updates face database.
Step 213, if K matching values of the face in face database are both less than second threshold, new personnel's name is established
It is single, using this K faces as newly-built face of the personnel in face database, the demographic data is added in face database.
Library algorithm is built in detector gate recognition of face and face the present invention is based on multiple-camera fusion automatically, belongs to non-formula
Face recognition algorithms and face build library algorithm, be obviously improved the speed of safety check, alleviated due to receiving to gather around caused by safety check
The security risks such as crowded, while the face database established automatically, complete data are provided for follow-up personal management.
Claims (8)
1. library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically, it is characterised in that:In detector gate
On in face of the direction that people enters multiple video cameras are installed, these video cameras acquire video and carry out Face datection, detect simultaneously
Face is screened by the laggard pedestrian's face quality assessment modules of face tracking, and the face after screening passes through the corresponding upper part of the body
The face for belonging to same person is selected with algorithm, then the excessive face of angle is deleted by Attitude estimation, is then aligned face
Depth convolutional neural networks extraction feature is inputted after calibration;Based on the face characteristic extracted above, pass through multiple-camera face
Fusion alignment algorithm carries out face matching and builds library automatically.
2. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that follow-on
Based on the Adaboost algorithm of Haar-like features, that is, the Haar-like features of 3x3 structures are increased, can preferably express people
The feature that face eyes, nose and mouth etc. are combined;Negative sample is from subway detector gate, airport security door, important field during training
The scenes such as shop detector gate are chosen, and are formed for the distinctive Face datection algorithm of detector gate scene.
3. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that are taken the photograph to each
The face of the appearance of camera is into line trace, and track algorithm uses the Kalman filter tracking algorithm based on neighborhood search, with target
Center is neighborhood search starting point, and by position and prediction of speed, detection range current face is nearest in a certain range of window
Candidate face, and pass through Kalman filter and be updated.
4. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that face quality
Assessment is carried out by the overall target to illuminance and clarity, and human face light degree evaluation index is by global average brightness, brightness
Peak, brightness minimum are calculated by formula, while add in illumination symmetry (negative and positive face) index;Face clarity refers to
Mark is counted by the number accounting of the high frequency coefficient after discrete cosine transform (DCT).
5. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that face characteristic
Vector is extracted by depth convolutional neural networks, which merges layer and 1 by 9 convolutional layers, 4 pond layers, 1
Full articulamentum composition;Convolutional layer uses the convolution kernel of 3x3, and pond layer uses the window of 2x2, merges layer by the 11st and the 12nd layer
Fusion Features are exported to next layer;Every layer of the convolutional neural networks all carry out feature normalization;When the convolutional neural networks are trained
The face database after detector gate scene calibration is employed to be trained.
6. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that will image more
Machine acquires corresponding face feature vector and is compared respectively with the face feature vector in face database, if these faces are in people
There is more than or equal to first threshold matching value, then select the first matching and the highest people of matching value is as the final of identification in face library
Output, while the first matched this is captured face and is added in corresponding face database, update face database.
7. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that multiple-camera
It acquires corresponding face matching value in face database and is both less than first threshold, then set up second threshold, more than or equal to second threshold,
Then think temporarily to match, every face in these faces selects most people (everyone of matching number from face database successively
Multiple corresponding storage faces), comprehensive score is calculated by formula, selects the corresponding people of comprehensive score highest as final face
It is added in corresponding face database with output, while the face of corresponding this candid photograph, updates face database.
8. detector gate recognition of face according to claim 1 and face build library algorithm automatically, which is characterized in that multiple-camera
It acquires matching value of the corresponding face in face database and is both less than second threshold, then new staff list is established, these faces
As face of the newly-built personnel in face database, the demographic data is added in face database.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810021107.2A CN108256459B (en) | 2018-01-10 | 2018-01-10 | Security check door face recognition and face automatic library building algorithm based on multi-camera fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810021107.2A CN108256459B (en) | 2018-01-10 | 2018-01-10 | Security check door face recognition and face automatic library building algorithm based on multi-camera fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108256459A true CN108256459A (en) | 2018-07-06 |
CN108256459B CN108256459B (en) | 2021-08-24 |
Family
ID=62726200
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810021107.2A Active CN108256459B (en) | 2018-01-10 | 2018-01-10 | Security check door face recognition and face automatic library building algorithm based on multi-camera fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108256459B (en) |
Cited By (39)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107977647A (en) * | 2017-12-20 | 2018-05-01 | 上海依图网络科技有限公司 | A kind of face recognition algorithms evaluating method of suitable public security actual combat |
CN108960216A (en) * | 2018-09-21 | 2018-12-07 | 浙江中正智能科技有限公司 | A kind of detection of dynamic human face and recognition methods |
CN109034013A (en) * | 2018-07-10 | 2018-12-18 | 腾讯科技(深圳)有限公司 | A kind of facial image recognition method, device and storage medium |
CN109190532A (en) * | 2018-08-21 | 2019-01-11 | 北京深瞐科技有限公司 | It is a kind of based on cloud side fusion face identification method, apparatus and system |
CN109271923A (en) * | 2018-09-14 | 2019-01-25 | 曜科智能科技(上海)有限公司 | Human face posture detection method, system, electric terminal and storage medium |
CN109376686A (en) * | 2018-11-14 | 2019-02-22 | 睿云联(厦门)网络通讯技术有限公司 | A kind of various dimensions human face data acquisition scheme, acquisition system and acquisition method |
CN109448848A (en) * | 2018-09-26 | 2019-03-08 | 长沙师范学院 | A kind of infantile psychology state evaluating method based on fuzzy evaluation |
CN109472247A (en) * | 2018-11-16 | 2019-03-15 | 西安电子科技大学 | Face identification method based on the non-formula of deep learning |
CN109472269A (en) * | 2018-10-17 | 2019-03-15 | 深圳壹账通智能科技有限公司 | Characteristics of image configuration and method of calibration, device, computer equipment and medium |
CN109508648A (en) * | 2018-10-22 | 2019-03-22 | 成都臻识科技发展有限公司 | A kind of face snap method and apparatus |
CN109615750A (en) * | 2018-12-29 | 2019-04-12 | 深圳市多度科技有限公司 | The recognition of face control method and device of door access machine, access control equipment, storage medium |
CN109635755A (en) * | 2018-12-17 | 2019-04-16 | 苏州市科远软件技术开发有限公司 | Face extraction method, apparatus and storage medium |
CN109685106A (en) * | 2018-11-19 | 2019-04-26 | 深圳博为教育科技有限公司 | A kind of image-recognizing method, face Work attendance method, device and system |
CN109711370A (en) * | 2018-12-29 | 2019-05-03 | 北京博睿视科技有限责任公司 | A kind of data anastomosing algorithm based on WIFI detection and face cluster |
CN109859085A (en) * | 2018-12-25 | 2019-06-07 | 深圳市天彦通信股份有限公司 | Safe early warning method and Related product |
CN109919091A (en) * | 2019-03-06 | 2019-06-21 | 广州佳都数据服务有限公司 | Face safety inspection method, device and electronic equipment based on dynamic white list |
CN110096965A (en) * | 2019-04-09 | 2019-08-06 | 华东师范大学 | A kind of face identification method based on head pose |
CN110321821A (en) * | 2019-06-24 | 2019-10-11 | 深圳爱莫科技有限公司 | Face alignment initial method and device, storage medium based on tripleplane |
CN110427864A (en) * | 2019-07-29 | 2019-11-08 | 腾讯科技(深圳)有限公司 | A kind of image processing method, device and electronic equipment |
CN110544333A (en) * | 2019-08-13 | 2019-12-06 | 成都电科慧安科技有限公司 | Access control system and control method thereof |
CN110825765A (en) * | 2019-10-23 | 2020-02-21 | 中国建设银行股份有限公司 | Face recognition method and device |
CN110908289A (en) * | 2018-09-17 | 2020-03-24 | 珠海格力电器股份有限公司 | Smart home control method and device |
CN111144366A (en) * | 2019-12-31 | 2020-05-12 | 中国电子科技集团公司信息科学研究院 | Strange face clustering method based on joint face quality assessment |
CN111310580A (en) * | 2020-01-19 | 2020-06-19 | 四川联众竞达科技有限公司 | Face recognition method under non-matching state |
CN111382592A (en) * | 2018-12-27 | 2020-07-07 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
CN111539351A (en) * | 2020-04-27 | 2020-08-14 | 广东电网有限责任公司广州供电局 | Multi-task cascaded face frame selection comparison method |
CN111652048A (en) * | 2020-04-17 | 2020-09-11 | 北京品恩科技股份有限公司 | A deep learning based 1: n face comparison method |
CN111814613A (en) * | 2020-06-24 | 2020-10-23 | 浙江大华技术股份有限公司 | Face recognition method, face recognition equipment and computer readable storage medium |
CN111866471A (en) * | 2020-07-31 | 2020-10-30 | 泽达易盛(天津)科技股份有限公司 | Visual intelligent security prevention and control terminal |
CN112001206A (en) * | 2019-05-27 | 2020-11-27 | 北京君正集成电路股份有限公司 | Method for merging face recognition libraries through traversal comparison |
CN112016508A (en) * | 2020-09-07 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Face recognition method, device, system, computing equipment and storage medium |
CN112132048A (en) * | 2020-09-24 | 2020-12-25 | 天津锋物科技有限公司 | Community patrol analysis method and system based on computer vision |
CN112307234A (en) * | 2020-11-03 | 2021-02-02 | 厦门兆慧网络科技有限公司 | Face bottom library synthesis method, system, device and storage medium |
CN113177530A (en) * | 2021-05-27 | 2021-07-27 | 广州广电运通智能科技有限公司 | Personnel screening method, equipment, medium and product |
CN113255608A (en) * | 2021-07-01 | 2021-08-13 | 杭州智爱时刻科技有限公司 | Multi-camera face recognition positioning method based on CNN classification |
CN113269124A (en) * | 2021-06-09 | 2021-08-17 | 重庆中科云从科技有限公司 | Object identification method, system, equipment and computer readable medium |
CN113688792A (en) * | 2021-09-22 | 2021-11-23 | 哈尔滨工程大学 | Face recognition method |
CN113779290A (en) * | 2021-09-01 | 2021-12-10 | 杭州视洞科技有限公司 | Camera face recognition aggregation optimization method |
CN116912808A (en) * | 2023-09-14 | 2023-10-20 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102306290A (en) * | 2011-10-14 | 2012-01-04 | 刘伟华 | Face tracking recognition technique based on video |
CN104978550A (en) * | 2014-04-08 | 2015-10-14 | 上海骏聿数码科技有限公司 | Face recognition method and system based on large-scale face database |
CN105654033A (en) * | 2015-12-21 | 2016-06-08 | 小米科技有限责任公司 | Face image verification 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 |
CN106156688A (en) * | 2015-03-10 | 2016-11-23 | 上海骏聿数码科技有限公司 | A kind of dynamic human face recognition methods and system |
CN106203260A (en) * | 2016-06-27 | 2016-12-07 | 南京邮电大学 | Pedestrian's recognition and tracking method based on multiple-camera monitoring network |
CN106295482A (en) * | 2015-06-11 | 2017-01-04 | ***(深圳)有限公司 | The update method of a kind of face database and device |
CN106845357A (en) * | 2016-12-26 | 2017-06-13 | 银江股份有限公司 | A kind of video human face detection and recognition methods based on multichannel network |
-
2018
- 2018-01-10 CN CN201810021107.2A patent/CN108256459B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102306290A (en) * | 2011-10-14 | 2012-01-04 | 刘伟华 | Face tracking recognition technique based on video |
CN104978550A (en) * | 2014-04-08 | 2015-10-14 | 上海骏聿数码科技有限公司 | Face recognition method and system based on large-scale face database |
CN106156688A (en) * | 2015-03-10 | 2016-11-23 | 上海骏聿数码科技有限公司 | A kind of dynamic human face recognition methods and system |
CN106295482A (en) * | 2015-06-11 | 2017-01-04 | ***(深圳)有限公司 | The update method of a kind of face database and device |
CN105654033A (en) * | 2015-12-21 | 2016-06-08 | 小米科技有限责任公司 | Face image verification 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 |
CN106203260A (en) * | 2016-06-27 | 2016-12-07 | 南京邮电大学 | Pedestrian's recognition and tracking method based on multiple-camera monitoring network |
CN106845357A (en) * | 2016-12-26 | 2017-06-13 | 银江股份有限公司 | A kind of video human face detection and recognition methods based on multichannel network |
Cited By (55)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107977647A (en) * | 2017-12-20 | 2018-05-01 | 上海依图网络科技有限公司 | A kind of face recognition algorithms evaluating method of suitable public security actual combat |
CN107977647B (en) * | 2017-12-20 | 2020-09-04 | 上海依图网络科技有限公司 | Face recognition algorithm evaluation method suitable for public security actual combat |
CN109034013A (en) * | 2018-07-10 | 2018-12-18 | 腾讯科技(深圳)有限公司 | A kind of facial image recognition method, device and storage medium |
CN109190532A (en) * | 2018-08-21 | 2019-01-11 | 北京深瞐科技有限公司 | It is a kind of based on cloud side fusion face identification method, apparatus and system |
CN109271923A (en) * | 2018-09-14 | 2019-01-25 | 曜科智能科技(上海)有限公司 | Human face posture detection method, system, electric terminal and storage medium |
CN110908289A (en) * | 2018-09-17 | 2020-03-24 | 珠海格力电器股份有限公司 | Smart home control method and device |
CN108960216A (en) * | 2018-09-21 | 2018-12-07 | 浙江中正智能科技有限公司 | A kind of detection of dynamic human face and recognition methods |
CN109448848A (en) * | 2018-09-26 | 2019-03-08 | 长沙师范学院 | A kind of infantile psychology state evaluating method based on fuzzy evaluation |
CN109472269A (en) * | 2018-10-17 | 2019-03-15 | 深圳壹账通智能科技有限公司 | Characteristics of image configuration and method of calibration, device, computer equipment and medium |
CN109508648A (en) * | 2018-10-22 | 2019-03-22 | 成都臻识科技发展有限公司 | A kind of face snap method and apparatus |
CN109376686A (en) * | 2018-11-14 | 2019-02-22 | 睿云联(厦门)网络通讯技术有限公司 | A kind of various dimensions human face data acquisition scheme, acquisition system and acquisition method |
CN109472247A (en) * | 2018-11-16 | 2019-03-15 | 西安电子科技大学 | Face identification method based on the non-formula of deep learning |
CN109472247B (en) * | 2018-11-16 | 2021-11-30 | 西安电子科技大学 | Face recognition method based on deep learning non-fit type |
CN109685106A (en) * | 2018-11-19 | 2019-04-26 | 深圳博为教育科技有限公司 | A kind of image-recognizing method, face Work attendance method, device and system |
CN109635755A (en) * | 2018-12-17 | 2019-04-16 | 苏州市科远软件技术开发有限公司 | Face extraction method, apparatus and storage medium |
CN109859085A (en) * | 2018-12-25 | 2019-06-07 | 深圳市天彦通信股份有限公司 | Safe early warning method and Related product |
CN111382592A (en) * | 2018-12-27 | 2020-07-07 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
CN111382592B (en) * | 2018-12-27 | 2023-09-29 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
US11682231B2 (en) | 2018-12-27 | 2023-06-20 | Hangzhou Hikvision Digital Technology Co., Ltd. | Living body detection method and device |
CN109615750A (en) * | 2018-12-29 | 2019-04-12 | 深圳市多度科技有限公司 | The recognition of face control method and device of door access machine, access control equipment, storage medium |
CN109711370B (en) * | 2018-12-29 | 2021-03-26 | 北京博睿视科技有限责任公司 | Data fusion method based on WIFI detection and face clustering |
CN109711370A (en) * | 2018-12-29 | 2019-05-03 | 北京博睿视科技有限责任公司 | A kind of data anastomosing algorithm based on WIFI detection and face cluster |
CN109919091A (en) * | 2019-03-06 | 2019-06-21 | 广州佳都数据服务有限公司 | Face safety inspection method, device and electronic equipment based on dynamic white list |
CN110096965A (en) * | 2019-04-09 | 2019-08-06 | 华东师范大学 | A kind of face identification method based on head pose |
CN112001206B (en) * | 2019-05-27 | 2023-09-22 | 北京君正集成电路股份有限公司 | Method for combining face recognition libraries through traversal comparison |
CN112001206A (en) * | 2019-05-27 | 2020-11-27 | 北京君正集成电路股份有限公司 | Method for merging face recognition libraries through traversal comparison |
CN110321821B (en) * | 2019-06-24 | 2022-10-25 | 深圳爱莫科技有限公司 | Human face alignment initialization method and device based on three-dimensional projection and storage medium |
CN110321821A (en) * | 2019-06-24 | 2019-10-11 | 深圳爱莫科技有限公司 | Face alignment initial method and device, storage medium based on tripleplane |
CN110427864A (en) * | 2019-07-29 | 2019-11-08 | 腾讯科技(深圳)有限公司 | A kind of image processing method, device and electronic equipment |
CN110427864B (en) * | 2019-07-29 | 2023-04-21 | 腾讯科技(深圳)有限公司 | Image processing method and device and electronic equipment |
CN110544333A (en) * | 2019-08-13 | 2019-12-06 | 成都电科慧安科技有限公司 | Access control system and control method thereof |
CN110544333B (en) * | 2019-08-13 | 2021-07-16 | 成都电科慧安科技有限公司 | Access control system and control method thereof |
CN110825765A (en) * | 2019-10-23 | 2020-02-21 | 中国建设银行股份有限公司 | Face recognition method and device |
CN110825765B (en) * | 2019-10-23 | 2022-10-04 | 中国建设银行股份有限公司 | Face recognition method and device |
CN111144366A (en) * | 2019-12-31 | 2020-05-12 | 中国电子科技集团公司信息科学研究院 | Strange face clustering method based on joint face quality assessment |
CN111310580A (en) * | 2020-01-19 | 2020-06-19 | 四川联众竞达科技有限公司 | Face recognition method under non-matching state |
CN111652048A (en) * | 2020-04-17 | 2020-09-11 | 北京品恩科技股份有限公司 | A deep learning based 1: n face comparison method |
CN111539351B (en) * | 2020-04-27 | 2023-11-03 | 广东电网有限责任公司广州供电局 | Multi-task cascading face frame selection comparison method |
CN111539351A (en) * | 2020-04-27 | 2020-08-14 | 广东电网有限责任公司广州供电局 | Multi-task cascaded face frame selection comparison method |
CN111814613A (en) * | 2020-06-24 | 2020-10-23 | 浙江大华技术股份有限公司 | Face recognition method, face recognition equipment and computer readable storage medium |
CN111866471B (en) * | 2020-07-31 | 2022-05-03 | 泽达易盛(天津)科技股份有限公司 | Visual intelligent public security prevention and control terminal |
CN111866471A (en) * | 2020-07-31 | 2020-10-30 | 泽达易盛(天津)科技股份有限公司 | Visual intelligent security prevention and control terminal |
CN112016508A (en) * | 2020-09-07 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Face recognition method, device, system, computing equipment and storage medium |
CN112016508B (en) * | 2020-09-07 | 2023-08-29 | 杭州海康威视数字技术股份有限公司 | Face recognition method, device, system, computing device and storage medium |
CN112132048A (en) * | 2020-09-24 | 2020-12-25 | 天津锋物科技有限公司 | Community patrol analysis method and system based on computer vision |
CN112307234A (en) * | 2020-11-03 | 2021-02-02 | 厦门兆慧网络科技有限公司 | Face bottom library synthesis method, system, device and storage medium |
CN113177530A (en) * | 2021-05-27 | 2021-07-27 | 广州广电运通智能科技有限公司 | Personnel screening method, equipment, medium and product |
CN113269124B (en) * | 2021-06-09 | 2023-05-09 | 重庆中科云从科技有限公司 | Object recognition method, system, equipment and computer readable medium |
CN113269124A (en) * | 2021-06-09 | 2021-08-17 | 重庆中科云从科技有限公司 | Object identification method, system, equipment and computer readable medium |
CN113255608A (en) * | 2021-07-01 | 2021-08-13 | 杭州智爱时刻科技有限公司 | Multi-camera face recognition positioning method based on CNN classification |
CN113779290A (en) * | 2021-09-01 | 2021-12-10 | 杭州视洞科技有限公司 | Camera face recognition aggregation optimization method |
CN113688792A (en) * | 2021-09-22 | 2021-11-23 | 哈尔滨工程大学 | Face recognition method |
CN113688792B (en) * | 2021-09-22 | 2023-12-08 | 哈尔滨工程大学 | Face recognition method |
CN116912808A (en) * | 2023-09-14 | 2023-10-20 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
CN116912808B (en) * | 2023-09-14 | 2023-12-01 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
Also Published As
Publication number | Publication date |
---|---|
CN108256459B (en) | 2021-08-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108256459A (en) | Library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically | |
CN109819208B (en) | Intensive population security monitoring management method based on artificial intelligence dynamic monitoring | |
CN109934176B (en) | Pedestrian recognition system, recognition method, and computer-readable storage medium | |
Shao et al. | Deep convolutional dynamic texture learning with adaptive channel-discriminability for 3D mask face anti-spoofing | |
CN104361327B (en) | A kind of pedestrian detection method and system | |
CN103106393B (en) | A kind of embedded human face identification intelligent identity authorization system based on robot platform | |
CN108416336B (en) | A kind of method and system of intelligence community recognition of face | |
CN110070010A (en) | A kind of face character correlating method identified again based on pedestrian | |
CN104166841B (en) | The quick detection recognition methods of pedestrian or vehicle is specified in a kind of video surveillance network | |
CN104091176B (en) | Portrait comparison application technology in video | |
CN109101871A (en) | A kind of living body detection device based on depth and Near Infrared Information, detection method and its application | |
CN104036236B (en) | A kind of face gender identification method based on multiparameter exponential weighting | |
CN103605971B (en) | Method and device for capturing face images | |
CN109858439A (en) | A kind of biopsy method and device based on face | |
CN110414441B (en) | Pedestrian track analysis method and system | |
CN103996033B (en) | Human identification method based on tracking of human face five-sense-organ coordinates | |
CN109255319A (en) | For the recognition of face payment information method for anti-counterfeit of still photo | |
CN113269091A (en) | Personnel trajectory analysis method, equipment and medium for intelligent park | |
RU2315352C2 (en) | Method and system for automatically finding three-dimensional images | |
CN114581990A (en) | Intelligent running test method and device | |
Denman et al. | Can you describe him for me? a technique for semantic person search in video | |
Islam et al. | Correlating belongings with passengers in a simulated airport security checkpoint | |
CN109858308B (en) | Video retrieval device, video retrieval method, and storage medium | |
Khanam et al. | Baggage recognition in occluded environment using boosting technique | |
Jeny et al. | Deep Learning Framework for Face Mask Detection |
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 | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20231125 Address after: Room 609-1, 6th Floor, Import and Export Exhibition and Trading Center, Huanghua Comprehensive Bonded Zone, Huanghua Town, Lingkong Block, Changsha Area, Changsha Free Trade Zone, Hunan Province, 410137 Patentee after: Hunan Shengxun Technology Co.,Ltd. Address before: Room 403, 4th Floor, Building 6, No. 13 North Ertiao, Zhongguancun, Haidian District, Beijing, 100190 Patentee before: BEIJING BRAVEVIDEO TECHNOLOGY CO.,LTD. |