CN109949452A - Channel based on recognition of face enters and leaves detection method - Google Patents
Channel based on recognition of face enters and leaves detection method Download PDFInfo
- Publication number
- CN109949452A CN109949452A CN201910130618.2A CN201910130618A CN109949452A CN 109949452 A CN109949452 A CN 109949452A CN 201910130618 A CN201910130618 A CN 201910130618A CN 109949452 A CN109949452 A CN 109949452A
- Authority
- CN
- China
- Prior art keywords
- initlevel
- face
- button
- picture
- recognition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of, and the channel based on recognition of face enters and leaves detection method, comprising: draws dialog box；Inquire InitLevel；The initialization progress of display system；Judge whether InitLevel=4, if so, killing timer, terminates inquiry；Otherwise, it returns；System queries message queue；Into information acquisition module；Into real-time monitoring module；It exits；Start additional initial work person's thread；InitLevel=1；It checks that Portraits file whether there is, is then created if it does not exist；InitLevel=2；Initialize Python environment；IInitLevel=3；Load FaceNet Model；InitLevel=4；Exit thread.The present invention realizes that the face recognition application under special screne, the government department to supervise by force provide preferable service experience, are that enterprise increases a set of competitive solution.
Description
Technical field
The present invention relates to Air conduct measurement field, in particular to a kind of channel based on recognition of face enters and leaves detection method.
Background technique
The technology of recognition of face is maked rapid progress now, and various application scenarios are seen everywhere, but are changed ten thousand times without leaving the original aim or stand, mainly
It is targetedly to be extracted as a kind of bio-identification means to interested information, and bring relevant information system into
In utilized.Various face recognition technologies are all the channels applied masses gather occasion or personnel pass through, in particular field
The application of conjunction is not common at present.Such as: it does not apply also in the meeting channel for meeting with process with family members to the personnel under detention of supervision place
Recognition of face detection technique.
Summary of the invention
The technical problem to be solved in the present invention is that in view of the above drawbacks of the prior art, providing a kind of realization particular field
Face recognition application under scape, the government department to supervise by force provide preferable service experience, are that enterprise's increase by one is cased with competition
The channel based on recognition of face of the solution of power enters and leaves detection method.
The technical solution adopted by the present invention to solve the technical problems is: constructing a kind of channel discrepancy based on recognition of face
Detection method includes the following steps:
A dialog box) is initialized；
B submodule block entrance button) is disabled；
C) draw dialog box, execute step D), D') or D ")；
D InitLevel) is inquired；
E) according to the initialization progress of the value display system of the InitLevel；
F) judge whether to meet InitLevel=4, if so, killing timer, terminate inquiry, execute step G)；Otherwise,
Return step D)；
G) the open submodule block entrance button；
D') system queries message queue executes step E'), F') or G')；
E' it) clicks information collection and enters information acquisition module；
F' it) clicks real time monitoring and enters real-time monitoring module；
G' it) clicks and is exited；
D ") the additional initial work person's thread of starting；
E ") InitLevel=1；
F ") check that Portraits file whether there is, it is then created if it does not exist；
G ") InitLevel=2；
H ") initialization Python environment；
I ") InitLevel=3；
J ") load FaceNet Model；
K ") InitLevel=4；
L ") exit thread.
Entered and left in detection method in the channel of the present invention based on recognition of face, the step E') further comprise:
E1') enter information acquisition module；
E2') dialog box window initializes；
E3' dialog box window) is drawn, step E4' is executed) or E4 ")；
E4' a frame picture) is extracted from camera；
E5') judge whether the picture is sky, if so, return step E4')；Otherwise, detection function is executed, step is executed
E6')；
E6') bounding box is plotted on the picture and shows the return step E4' on IDC_VIDEO control)；
E4 ") systems inspection message queue, execute step E5 "), E6 ") or E7 ")；
E5 ") if catching press-button is clicked, enter catching press-button and clicks receptance function；
E6 ") if it is determined that button is clicked, then enter confirming button and clicks receptance function；
E7 ") if exit button is clicked, return to main interface.
It is entered and left in detection method in the channel of the present invention based on recognition of face, the catching press-button clicks response letter
Several process flows include:
E51 ") a frame picture is extracted from camera；
E52 ") execute detection function；
E53 ") judge whether to detect face, if so, executing step E55 ")；Otherwise, step E54 " is executed)；
E54 ") corresponding prompt information is provided, execute step E59 ")；
E55 ") judge whether face's quantity is 1, if so, executing step E56 ")；Otherwise, return step E54 ")；
E56 ") judge whether the size of face is suitable, if so, cutting out and scaling picture to 400 × 400, execution step
E57")；Otherwise, return step E54 ")；
E57 ") save the picture and normalized picture captured；
E58 ") 400 × 400 normalization picture is shown on IDC_SHOW control, execute step E59 ")；
E59 ") terminate.
It is entered and left in detection method in the channel of the present invention based on recognition of face, the confirming button clicks response letter
Several process flows include:
E61 ") judge whether to capture portrait, if so, executing step E63 ")；Otherwise, step E62 " is executed)；
E62 ") output prompt information, execute step E69 ")；
E63 ") judge whether to input name, if so, creating a people object Person, execute step E64 ")；It is no
Then, return step E62 ")；
E64 ") take out the picture captured and testing result；
E65 ") Name, PortaritInput and PortraitShow of the Person are set；
E66 ") calculate the mappings characteristics of the Person；
E67 ") 400 × 400 normalized personage's head portraits are stored in file system；
E68 ") database is written in the information of Person described in part, execute step E69 ")；
E69 ") terminate.
It is entered and left in detection method in the channel of the present invention based on recognition of face, the database is Face datection
Library.
Implement the channel of the invention based on recognition of face and enter and leave detection method, has the advantages that due to executing
Following steps: A) initialization dialog box；B submodule block entrance button) is disabled；C) draw dialog box, execute step D), D') or
D")；D InitLevel) is inquired；E) according to the initialization progress of the value display system of the InitLevel；F) judge whether full
Sufficient InitLevel=4 terminates inquiry if so, killing timer, executes step G)；Otherwise, return step D)；G) described in open
Submodule block entrance button；D') system queries message queue executes step E'), F') or G')；E' it) clicks information collection and enters letter
Cease acquisition module；F' it) clicks real time monitoring and enters real-time monitoring module；G' it) clicks and is exited；D ") start additionally
Initial work person's thread；E ") InitLevel=1；F ") check Portraits file whether there is, if it does not exist then into
Row creation；G ") InitLevel=2；H ") initialization Python environment；I ") InitLevel=3；J ") load FaceNet
Model；K ") InitLevel=4；L ") exit thread；Because prison system is a special industry, some of commercialization are disclosed
Product or equipment, are not directly applicable in the actual demand of supervision place, therefore are directed to some special requirements, the present invention
It proposes that a kind of meeting channel based on recognition of face enters and leaves detection method, can be applied to prison system, realize special screne
Under face recognition application, the government department to supervise by force provide preferable service experience, be enterprise increase it is a set of competitive
Solution.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is to enter and leave in detection method one embodiment to be directed toward storage testing result the present invention is based on the channel of recognition of face
The data structure schematic diagram of the pointer in region；
Fig. 2 is the schematic diagram of Face datection result in the embodiment；
Fig. 3 is the schematic diagram of data flow and conversion process in the embodiment.
Fig. 4 is the flow chart of the channel discrepancy detection method in the embodiment based on recognition of face；
Fig. 5 is to click the specific flow chart that information collection enters information acquisition module in the embodiment；
Fig. 6 is the process flow diagram that catching press-button clicks receptance function in the embodiment；
Fig. 7 is the process flow diagram that confirming button clicks receptance function in the embodiment.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The present invention is based in the channel of recognition of face discrepancy detection method embodiment, it is somebody's turn to do the channel based on recognition of face and goes out
Enter the control for the process that detection method is laid particular emphasis on to acquisition, characteristic extraction, identification and the comparison accuracy that need to identify object etc.,
Certain optimization has been done to testing process.
About the realization of detection part, the database in the present invention uses Face datection library, incites somebody to action after having configured the library dll
Correspondingly header file imported into item file the interface that this library can be used, which is directed to different applied fields
Conjunction provides four different interfaces, and the present invention uses detection effect best but a most time-consuming interface:
int*facedetect_multiview_reinforce(parameter list….)；
Function parameter list includes the pointer for being directed toward storage testing result region, is directed toward storage input image data region
Pointer, input the length and width and some other parameters of picture, function returns to the pointer for being directed toward storage testing result region,
This to storage testing result region pointer data structure schematic diagram it is as shown in Figure 1.Key seeks to extract this pointer institute
The data of direction, it is assumed that the pointer of return is pResult, then first integer data that pResult is directed toward, which is meant that, to be inputted
The face quantity detected in picture houses everyone face information in the continuous region of memory that gets off, everyone
142 integer datas are accounted for, first four number indicates the upper left corner of face bounding box and the left side in the lower right corner, and the 5th data indicate name
For the information (its concrete meaning is also unknown at present) of neighbor, the 6th data indicate the corner of face, and range is 0~60
Degree, every two data form a coordinate points later, mark 68 characteristic points of face, these characteristic points mark the wheel of face
Exterior feature, eyes, eyebrow, nose, the position of mouth and profile occupy 136 data, the schematic diagram of Face datection result such as Fig. 2 altogether
It is shown.
Detection based on monitoring only needs to be spaced from camera at regular intervals to take out picture, and picture is sent into and is detected
Detection structure is plotted on picture on being output to user interface after function.
About the realization for extracting identification feature part, the present invention is used will be used based on Tensorflow reality Cython
The function that Python is defined is converted to the function that C-shaped formula defines, and a FaceNet neural network model of C-shaped formula uses Python
The function for directly using Tensorflow is write, the header file and source file of respective function can be automatically generated by ython, it will
The function that C-shaped formula can be used to define in the file lead-in item of generation.The use of all about Tensorflow is all packed
In a class for being TensorflowInference, acquiring and maintaining behaviour of this class wrapper including session object
Work, the load of network model, calculates mappings characteristics and all C++ and Python data friendships at the initialization of Python environment
Interworking, to class outside user conceal the content of related to Python completely.Thering is two o'clock to need to pay attention among these:
First point is Session object, and calculating in Tensorflow all is completed by session object,
The structural information and value information of data flow diagram are saved in session object, the network model that the present invention uses is very huge
Model is all called in memory from disk again if calculated every time, can seriously increase the time-consuming (load every time of program by (141MB)
Model is two minutes about time-consuming), thus reasonable design should be model is just loaded into memory in whole system initialization, and
Obtaining its session object makes its memory-resident, calls the member method of session object to can be completed once when calculating every time
It calculates and (calculates every time about 0.3~0.4 second time-consuming).In Python script, the acquisition of session object is to pass through
What the interface tf.Session () that Tensorflow is provided was obtained, but before obtaining session object, it needs data
The structural information of flow graph is loaded into the graph of default, then the weight in figure is restored in session object, in this way after
Session object could correctly execute the forward calculation of FaceNet.The session object of acquisition can be in C++ with PyObject
Type occur.
Second point is the data interaction of C++ and Python, and a difficult point is exactly the data interaction problem of C++ and Python,
The form of data interaction is carried out by the parameter and return value of function, includes mainly both direction, first is that from the part C++
Data flow into Python, second is that the part Python, which executes the result calculated, returns to C++, are according to the way that Python official provides
The data conversion of all C++ class types is interacted at PyObject object, but since Python all objects are in C++
It is entirely inherited from PyObject type, so flexible data type brings certain complexity, and makes in the present invention
It is numpy.ndarray type in Python with picture type data, numpy official gives to C++ version
Ndarray supports, but when showing that the function interface used has been subjected in actual use.Fortunately Cython allows in Python
The data structure that C type is used in script, so that the data conversion from C++ to Python facilitates much.Specific data flow
To as shown in Figure 3 with conversion process.
The major function of main interface is the Python running environment of initialization system, loads FaceNet Model, and show
The entrance of information acquisition module and real-time monitoring module.The channel based on recognition of face enters and leaves the flow chart of detection method such as
Shown in Fig. 4, in Fig. 4, it includes following step that detection method (i.e. the execution process of main interface) should be entered and left based on the channel of recognition of face
It is rapid:
Step S01 initializes dialog box: in this step, initializing dialog box.
Step S02 disables submodule block entrance button: in this step, disabling submodule block entrance button.
Step S03 draws dialog box: in this step, drawing dialog box.Executed this step, execute step S04, S04' or
S04"。
Step S04 inquires InitLevel: in this step, inquiring InitLevel, has executed this step, executes step S05.
Step S05 is according to the initialization progress of the value display system of InitLevel: in this step, according to InitLevel's
It is worth the initialization progress of display system.This step has been executed, step S06 is executed.
Step S06 judges whether to meet InitLevel=4: in this step, judging whether to meet InitLevel=4, such as
The result that fruit judges be it is yes, then follow the steps S07；Otherwise, return step S04.
Step S07 kills timer, terminates inquiry: in this step, killing timer, terminates inquiry.This step has been executed,
Execute step S08.
Step S08 opens submodule block entrance button: in this step, open submodule block entrance button.
Step S04' system queries message queue: in this step, system queries message queue.This step has been executed, has been executed
Step S05', S06' or S07'.
Step S05' clicks information collection and enters information acquisition module: in this step, clicking information collection and enters information and adopt
Collect module.
Step S06' clicks real time monitoring and enters real-time monitoring module: in this step, clicking real time monitoring and enters prison in real time
Control module.
Step S07' is clicked and is exited: in this step, being clicked and is exited.
Step S04 " starts additional initial work person's thread: in this step, starting additional initial work person's line
Journey.
Step S05 " InitLevel=1: in this step, InitLevel=1.
Step S06 " checks that Portraits file whether there is, and is then created if it does not exist: in this step, checking
Portraits file whether there is, and then be created if it does not exist.
Step S07 " InitLevel=2: in this step, InitLevel=2.
Step S08 " initializes Python environment: in this step, initializing Python environment.
Step S09 " InitLevel=3: in this step, InitLevel=3.
Step S10 " loads FaceNet Model: in this step, loading FaceNet Model.
Step S11 " InitLevel=4: in this step, InitLevel=4.
Step S12 " exits thread: in this step, exiting thread.
It is noted that having rewritten following function in the class that VS is main interface generation in the present embodiment:
OnInitDialog function: for executing simple not time-consuming initial work, if being executed in this function
Time-consuming initial work, will lead to that window is stuck before calling drafting function, and a kind of design of compromise is at the beginning of time-consuming
Beginning chemical industry is placed in OnPaint () function, executes time-consuming initial work again after window is all drawn.This hair
The Cheng Zhihang although the bright initial work by time-consuming separately bursts at the seams, if starting in OnInitDialog function by test discovery
Thread is also possible to will lead to forms and can not normally draw.
OnPaint function: start additional initialization thread in the ending of this function, it is ensured that finish in forms drafting
Time-consuming initial work is executed afterwards.It is to be noted that can also be called if user drags window or changes the size of window
Whether the function is completed in system with the initialization of the variable label of a BOOL type in order to avoid repeating for initialization,
Only just start additional initialization thread in the case where unfinished, once initialization is completed, just at once by this BOOL class
The variable of type negates, thus can be to avoid starting additional initialization thread in drafting later.
OnTimer function: system can start at the SetTimer of program, just call this function every a bit of time,
Its critical function is the value of continuous inquiry InitLevel, and initialization progress is shown to user according to this value.
In the present embodiment, the information collection button by clicking main interface enters information acquisition module, his function is to receive
Collect the information of personage and database is written.
For the present embodiment, above-mentioned steps S05' can also be refined further, flow chart such as Fig. 5 institute after refinement
Show.In Fig. 5, step S05' further comprises following steps:
Step S51' enters information acquisition module: in this step, into information acquisition module.
The initialization of step S52' dialog box window: in this step, dialog box window initialization.
Step S53' draws dialog box window: in this step, drawing dialog box window, has executed this step, executes step
S54' or step S54 ".
Step S54' is from extracting a frame picture in camera: in this step, a frame picture is extracted from camera.
Step S55' judges whether picture is empty: in this step, judge whether picture is empty, if it is determined that result be
It is, then return step S54'；Otherwise, step S56' is executed.
Step S56' executes detection function: in this step, executing detection function.This step has been executed, step S57' is executed.
Bounding box is plotted on picture and shows on IDC_VIDEO control by step S57': in this step, by bounding box
It is plotted on picture and is shown on IDC_VIDEO control.
Step S54 " systems inspection message queue: in this step, systems inspection message queue.This step has been executed, has been executed
Step S55 ", S56 " or step S57 ".
Step S55 " enters catching press-button and clicks receptance function if catching press-button is clicked: in this step, if capture is pressed
Button is clicked, then enters catching press-button and click receptance function.
Step S56 " then enters confirming button and clicks receptance function: in this step, however, it is determined that press if it is determined that button is clicked
Button is clicked, then enters confirming button and click receptance function.
Step S57 " returns to main interface if exit button is clicked: in this step, if exit button is clicked, returning
Return main interface.The dialog class that the realization of information acquisition module uses a MFC to generate, message response function is all conduct
The member function of this class is realized.
Fig. 6 is that catching press-button clicks the process flow diagram of receptance function in the present embodiment, and in Fig. 6, catching press-button clicks sound
The process flow for answering function CInfoCollectDlg::OnBnClickedButtonCaptrue () further comprises walking as follows
It is rapid:
Step S551 " is from extracting a frame picture in camera: in this step, a frame picture is extracted from camera.
Step S552 " executes detection function: in this step, executing detection function.
Step S553 " judges whether to detect face: in this step, judge whether to detect face, if it is determined that knot
Fruit be it is yes, then follow the steps S555 "；Otherwise, step S554 " is executed.
Step S554 " provides corresponding prompt information: in this step, providing corresponding prompt information.This step has been executed,
Execute step S560 ".
Step S555 " judges whether face's quantity is 1: in this step, judge whether face's quantity is 1, if it is determined that
It as a result is yes, execution step S556 "；Otherwise, step S554 " is executed.
Step S556 " judges whether the size of face is suitable: in this step, judge whether the size of face is suitable, if
The result judged be it is yes, then follow the steps S557 "；Otherwise, step S554 " is executed.
Step S557 " is cut out with scaling picture to 400 × 400: in this step, being cut out and scaling picture to 400 × 400.
Step S558 " saves the picture and normalized picture captured: in this step, saving the picture captured and rule
The picture formatted.
400 × 400 normalization picture is shown on IDC_SHOW control by step S559 ": in this step, by 400 ×
400 normalization picture is shown on IDC_SHOW control.
Step S560 " terminates: in this step, terminating.
Fig. 7 is that confirming button clicks the process flow diagram of receptance function in the present embodiment, in Fig. 7, the message of confirming button
The process flow of receptance function CInfoCollectDlg::OnBnClickedButtonComfier () includes the following steps:
Step S561 " judges whether to capture portrait: in this step, judge whether to capture portrait, if it is determined that knot
Fruit be it is yes, then follow the steps S562 "；Otherwise, step S563 " is executed.
Step S562 " exports prompt information: in this step, exporting prompt information.This step has been executed, step is executed
S570"。
Step S563 " judge whether input name: in this step, judge whether input name, if it is determined that result be
It is to then follow the steps S564 "；Otherwise, step S562 " is executed.
Step S564 " creates a people object Person: in this step, creating a people object Person.
Step S565 " takes out the picture captured and testing result: in this step, taking out the picture captured and detection knot
Fruit.
Name, PortaritInput and PortraitShow of step S566 " setting Person: in this step, setting
Name, PortaritInput and PortraitShow of Person.
The mappings characteristics of step S567 " calculating Person: in this step, the mappings characteristics of Person are calculated.
400 × 400 normalized personage's head portraits are stored in file system by step S568 ": in this step, by 400 × 400 rule
The personage's head portrait deposit file system formatted.
Database is written in the information of Person by step S569 ": in this step, database is written in the information of Person.
Step S570 " terminates: in this step, terminating.
In short, the present invention provides a good try for the face recognition application under special screne, also to supervise by force
Government department provide primary preferable service experience.A set of competitive solution is increased for enterprise, it is pre- in the future
Meter has preferable income.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (5)
1. a kind of channel based on recognition of face enters and leaves detection method, which comprises the steps of:
A dialog box) is initialized；
B submodule block entrance button) is disabled；
C) draw dialog box, execute step D), D') or D ")；
D InitLevel) is inquired；
E) according to the initialization progress of the value display system of the InitLevel；
F) judge whether to meet InitLevel=4, if so, killing timer, terminate inquiry, execute step G)；Otherwise, it returns
Step D)；
G) the open submodule block entrance button；
D') system queries message queue executes step E'), F') or G')；
E' it) clicks information collection and enters information acquisition module；
F' it) clicks real time monitoring and enters real-time monitoring module；
G' it) clicks and is exited；
D ") the additional initial work person's thread of starting；
E ") InitLevel=1；
F ") check that Portraits file whether there is, it is then created if it does not exist；
G ") InitLevel=2；
H ") initialization Python environment；
I ") InitLevel=3；
J ") load FaceNet Model；
K ") InitLevel=4；
L ") exit thread.
2. the channel according to claim 1 based on recognition of face enters and leaves detection method, which is characterized in that the step
E') further comprise:
E1') enter information acquisition module；
E2') dialog box window initializes；
E3' dialog box window) is drawn, step E4' is executed) or E4 ")；
E4' a frame picture) is extracted from camera；
E5') judge whether the picture is sky, if so, return step E4')；Otherwise, detection function is executed, step is executed
E6')；
E6') bounding box is plotted on the picture and shows the return step E4' on IDC_VIDEO control)；
E4 ") systems inspection message queue, execute step E5 "), E6 ") or E7 ")；
E5 ") if catching press-button is clicked, enter catching press-button and clicks receptance function；
E6 ") if it is determined that button is clicked, then enter confirming button and clicks receptance function；
E7 ") if exit button is clicked, return to main interface.
3. the channel according to claim 2 based on recognition of face enters and leaves detection method, which is characterized in that the capture is pressed
The process flow that button clicks receptance function includes:
E51 ") a frame picture is extracted from camera；
E52 ") execute detection function；
E53 ") judge whether to detect face, if so, executing step E55 ")；Otherwise, step E54 " is executed)；
E54 ") corresponding prompt information is provided, execute step E59 ")；
E55 ") judge whether face's quantity is 1, if so, executing step E56 ")；Otherwise, return step E54 ")；
E56 ") judge whether the size of face is suitable, if so, cutting out and scaling picture to 400 × 400, execution step E57 ")；
Otherwise, return step E54 ")；
E57 ") save the picture and normalized picture captured；
E58 ") 400 × 400 normalization picture is shown on IDC_SHOW control, execute step E59 ")；
E59 ") terminate.
4. the channel according to claim 3 based on recognition of face enters and leaves detection method, which is characterized in that the determination is pressed
The process flow that button clicks receptance function includes:
E61 ") judge whether to capture portrait, if so, executing step E63 ")；Otherwise, step E62 " is executed)；
E62 ") output prompt information, execute step E69 ")；
E63 ") judge whether to input name, if so, creating a people object Person, execute step E64 ")；Otherwise, it returns
Return step E62 ")；
E64 ") take out the picture captured and testing result；
E65 ") Name, PortaritInput and PortraitShow of the Person are set；
E66 ") calculate the mappings characteristics of the Person；
E67 ") 400 × 400 normalized personage's head portraits are stored in file system；
E68 ") by the information write-in database of the Person, execute step E69 ")；
E69 ") terminate.
5. the channel according to claim 4 based on recognition of face enters and leaves detection method, which is characterized in that the database
For Face datection library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910130618.2A CN109949452A (en) | 2019-02-21 | 2019-02-21 | Channel based on recognition of face enters and leaves detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910130618.2A CN109949452A (en) | 2019-02-21 | 2019-02-21 | Channel based on recognition of face enters and leaves detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109949452A true CN109949452A (en) | 2019-06-28 |
Family
ID=67006875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910130618.2A Pending CN109949452A (en) | 2019-02-21 | 2019-02-21 | Channel based on recognition of face enters and leaves detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949452A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060289618A1 (en) * | 2005-06-22 | 2006-12-28 | Fuji Photo Film Co., Ltd. | Apparatus and method for deciding in-focus position of imaging lens |
| CN106203533A (en) * | 2016-07-26 | 2016-12-07 | 厦门大学 | The degree of depth based on combined training study face verification method |
| CN106648743A (en) * | 2016-10-19 | 2017-05-10 | 广州视源电子科技股份有限公司 | Initialization method and device for terminal |
| CN106873485A (en) * | 2017-02-28 | 2017-06-20 | 安徽梵凯思数码科技有限公司 | One kind building intelligent building monitoring system |
| CN106960185A (en) * | 2017-03-10 | 2017-07-18 | 陕西师范大学 | The Pose-varied face recognition method of linear discriminant depth belief network |
-
2019
- 2019-02-21 CN CN201910130618.2A patent/CN109949452A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060289618A1 (en) * | 2005-06-22 | 2006-12-28 | Fuji Photo Film Co., Ltd. | Apparatus and method for deciding in-focus position of imaging lens |
| CN106203533A (en) * | 2016-07-26 | 2016-12-07 | 厦门大学 | The degree of depth based on combined training study face verification method |
| CN106648743A (en) * | 2016-10-19 | 2017-05-10 | 广州视源电子科技股份有限公司 | Initialization method and device for terminal |
| CN106873485A (en) * | 2017-02-28 | 2017-06-20 | 安徽梵凯思数码科技有限公司 | One kind building intelligent building monitoring system |
| CN106960185A (en) * | 2017-03-10 | 2017-07-18 | 陕西师范大学 | The Pose-varied face recognition method of linear discriminant depth belief network |
Non-Patent Citations (1)
| Title |
|---|
| 李骏驰: "基于人脸识别的公司考勤门禁系统的设计", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105205646A (en) | Automatic roll call system and realization method thereof | |
| CN107481327B (en) | About the processing method of augmented reality scene, device, terminal device and system | |
| EP3767522A1 (en) | Image recognition method and apparatus, and terminal and storage medium | |
| CN109658928A (en) | A kind of home-services robot cloud multi-modal dialog method, apparatus and system | |
| CN113365147B (en) | Video editing method, device, equipment and storage medium based on music card point | |
| CN109815795A (en) | Classroom student's state analysis method and device based on face monitoring | |
| CN107360157A (en) | A kind of user registering method, device and intelligent air conditioner | |
| CN105160318A (en) | Facial expression based lie detection method and system | |
| CN106778496A (en) | Biopsy method and device | |
| CN113111690B (en) | Facial expression analysis method and system and satisfaction analysis method and system | |
| CN109447156A (en) | Method and apparatus for generating model | |
| CN110309799A (en) | Judgment method of speaking based on camera | |
| CN106710020A (en) | Intelligent attendance checking method and system | |
| CN109446946A (en) | A kind of multi-cam real-time detection method based on multithreading | |
| CN105468161A (en) | Instruction execution method and device | |
| CN110956059A (en) | Dynamic gesture recognition method and device and electronic equipment | |
| CN111191506A (en) | Personnel flow statistical method and device, computer equipment and storage medium | |
| CN111881320A (en) | Video query method, device, equipment and readable storage medium | |
| CN109710733A (en) | A kind of data interactive method and system based on intelligent sound identification | |
| CN110399810B (en) | Auxiliary roll-call method and device | |
| KR20200092166A (en) | Server, method and computer program for recognizing emotion | |
| CN109345427B (en) | Classroom video frequency point arrival method combining face recognition technology and pedestrian recognition technology | |
| CN109949452A (en) | Channel based on recognition of face enters and leaves detection method | |
| CN109857829A (en) | A kind of geographic information data fusion system | |
| CN114220175B (en) | Motion pattern recognition method and device, equipment, medium and product thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190628 |
|
| RJ01 | Rejection of invention patent application after publication |