CN108596171A - Enabling control method and system - Google Patents
Enabling control method and system Download PDFInfo
- Publication number
- CN108596171A CN108596171A CN201810272557.9A CN201810272557A CN108596171A CN 108596171 A CN108596171 A CN 108596171A CN 201810272557 A CN201810272557 A CN 201810272557A CN 108596171 A CN108596171 A CN 108596171A
- Authority
- CN
- China
- Prior art keywords
- human body
- children
- neural networks
- convolutional neural
- enabling control
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/30—Individual registration on entry or exit not involving the use of a pass
- G07C9/32—Individual registration on entry or exit not involving the use of a pass in combination with an identity check
- G07C9/37—Individual registration on entry or exit not involving the use of a pass in combination with an identity check using biometric data, e.g. fingerprints, iris scans or voice recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Multimedia (AREA)
- Lock And Its Accessories (AREA)
Abstract
The invention discloses a kind of enabling control method and systems, after acquisition human body region of interest area image, human body classification is carried out to the human body region of interest area image using convolutional neural networks model, door body is controlled when human body is classified as children and keeps locking, and door body is controlled when human body is classified as non-children and is unlocked；It is this using the body of entire people as area-of-interest, by way of judging classification age-related characteristic values such as apparel characteristic, figure and features feature, height feature, the fat or thin features of convolutional neural networks extraction people, so that classification results more closing to reality is more accurate, and shift to an earlier date typing authorization message without user, it is a kind of accurately and efficiently mode classification, it can prevent children from arbitrarily opening door body, avoid potential danger.
Description
Technical field
The invention belongs to image procossing identification technology fields, specifically, being to be related to a kind of enabling control method and system.
Background technology
Popularizing for cabinet, household electrical appliance with door body etc. is provided convenience for user's life, but there is also certain safety
Hidden danger；For example, filling up the cabinet of article, refrigerator etc. when opening door body, there are the possibility that article is fallen, if youngster at this time
Child opens door body, then the article that may be dropped accidentally injures head etc.；Alternatively, equal door body of opening pierces cabinet to children out of curiosity
In, other injury accidents as accidental cause can't get out or occur.
In order to avoid the potential injury to children, in the prior art, there are some enabling control methods, for example, with infrared
The height of human body occurred before sensor detection door body, when height determines threshold value less than a height, then at the lock that controls door body
In locking state, child is prevented, but relatively difficult for the setting of height threshold value in this scheme, it is right when setting relatively low
Door body can not be then opened in the shorter adult of height, setting is then unfavorable for protecting children when higher；In another example being obtained using camera
The face of user is identified in the mode for taking family facial information, and unauthorized user can not open door body or authorized user
Door body can be opened, but this mode needs the facial information of typing user, setting authorization message etc. in advance, occupying system resources
And it is inconvenient to use.
Invention content
This application provides a kind of enabling control method and systems, and it is accurate and efficient right can be based on user's human body information
User classifies, and prevents children from arbitrarily opening door body, avoids potential danger.
In order to solve the above technical problems, the application is achieved using following technical scheme：
It is proposed a kind of enabling control method, including：Obtain human body region of interest area image；Using convolutional neural networks model to institute
It states human body region of interest area image and carries out human body classification；Door body is controlled when the human body is classified as children and keeps locking, in institute
Control door body when human body is classified as non-children is stated to unlock.
Further, the convolutional neural networks model includes three convolutional layers, two full articulamentums and a grader,
People's is classified as children and non-children corresponding to the grader output human body.
Further, the convolutional neural networks model can be based on the update of offline or online mode.
Further, the convolutional neural networks are trained using stochastic gradient descent method.
It proposes a kind of enabling control system, including door body, further includes camera unit, taxon and control unit；It is described
Camera unit, for obtaining human body region of interest area image；The taxon, for using convolutional neural networks module to institute
It states human body region of interest area image and carries out human body classification；Described control unit, for being controlled when the human body is classified as children
The door body keeps locking, and the door body is controlled when the human body is classified as non-children and is unlocked.
Further, the convolutional neural networks model includes three convolutional layers, two full articulamentums and a grader,
People's is classified as children and non-children corresponding to the grader output human body.
Further, the system also includes storage units and updating unit；The storage unit, for storing the volume
Product neural network model；The updating unit, for to the convolutional neural networks model be based on offline or online mode into
Row update.
Further, the convolutional neural networks are trained using stochastic gradient descent method.
Compared with prior art, the advantages of the application and good effect is：The enabling control method and be that the application proposes
In system, the human body region of interest area image of the user occurred before door body is obtained using camera unit, the human body of acquisition is interested
Area image inputs trained convolutional neural networks model and carries out human body classification, and classification output is children or non-children, in people
Door body is then controlled when body is classified as children and keeps locking, and be non-children is to unlock so that the crowd of the non-children mankind can be with
Force opens door body after being unlocked, this using the body of entire people as area-of-interest, extracts people's by convolutional neural networks
The age-related characteristic values such as apparel characteristic, figure and features feature, height feature, fat or thin feature come judge classification mode,
So that classification results more closing to reality is more accurate, and shifts to an earlier date typing authorization message without user, be one kind accurately and efficiently
Mode classification can prevent children from arbitrarily opening door body, avoid potential danger.
After the detailed description of the application embodiment is read in conjunction with the figure, other features and advantages of the application will become more
Add clear.
Description of the drawings
Fig. 1 is the flow chart for the enabling control method that the application proposes；
Fig. 2 is the system block diagram for the enabling control system that the application proposes.
Specific implementation mode
The specific implementation mode of the application is described in more detail below in conjunction with the accompanying drawings.
The enabling control method that the application proposes, as shown in Figure 1, including：
Step S11:Obtain human body region of interest area image.
Camera unit is nearby set in door body, door body of refrigerator of cabinet etc. so that camera unit can obtain people
Body-sensing interest area image, specifically, the pedestrian detection method based on Yolo, which may be used, obtains human body area-of-interest.
Step S12：Human body classification is carried out to the human body region of interest area image using convolutional neural networks model.
Human body region of interest area image is sent in trained convolutional neural networks model, specifically, convolutional Neural net
Network model includes three convolutional layers, two full articulamentums and a grader, and first layer convolutional layer uses 96 convolution kernels, each
Convolution kernel number of parameters is, convolution step-length is 4, and activation primitive uses Relu, pondization to use maximum value pond, pond
The size of change is selected, pond step-length stride selections 2；Second layer convolutional layer uses 256 convolution kernels, convolution kernel size
For, convolution step-length is 1, and the size in pond is selected, pond step-length stride selections 2；Third layer convolutional layer uses
384 filters, convolution kernel size are；First full articulamentum neuron number selection 512, second full articulamentum god
512 are also selected through first number, output layer is set as two：Children and the grader of non-children namely convolutional neural networks model
People's is classified as children or non-children corresponding to output human body.
For convolutional neural networks model in training, weights initialisation method use standard deviation for 0.01, mean value for 0 Gauss
Just it is distributed very much；Prevent over-fitting from using two kinds of strategies in training：1, using dropout methods, dropout ratios use 0.5；2、
Data extending, data extending is taken to pass through inputPicture, then carry out random cropping, be cut toPicture, cutting carried out based on human body center；Training method uses stochastic gradient descent method, mini-
Batch sizes selection 50, learning rate is adjusted to 0.0001 by learning rate size 0.001 then after iterating to 10000 times.
Step S13：Door body is controlled when human body is classified as children and keeps locking, and door is controlled when human body is classified as non-children
Body is unlocked.
When convolutional neural networks model output result is children, then door body locking is controlled, user can not open door body, if defeated
Go out result be non-children when, then control door body lock unlock so that user force after can open door body.
In the embodiment of the present application, convolutional neural networks model can be based on the update of offline or online mode.
In the enabling control method that above-mentioned the application proposes, the human body of the user occurred before door body is obtained using camera unit
The human body region of interest area image of acquisition is inputted trained convolutional neural networks model and carries out human body by region of interest area image
Classification, classification output is children or non-children, and door body is then controlled when human body is classified as children and keeps locking, and is non-children to be
It then unlocks so that the crowd of the non-children mankind can exert a force after being unlocked opens door body, and this body with entire people is to feel emerging
Interesting region extracts apparel characteristic, figure and features feature, height feature, fat or thin feature of people etc. and year by convolutional neural networks
Age relevant characteristic value come judge classification mode so that classification results more closing to reality is more accurate, and is carried without user
Preceding typing authorization message is a kind of accurately and efficiently mode classification, can prevent children from arbitrarily opening door body, avoid potential danger
Danger occurs.
Based on enabling control method set forth above, the application also proposes a kind of enabling control system, as shown in Fig. 2, packet
Include door body 21, camera unit 22, taxon 23 and control unit 24；Camera unit 22 is for obtaining the human body sense before door body 21
Interest area image；Taxon 23 is used to carry out human body point to human body region of interest area image using convolutional neural networks module
Class；Control unit 24 is used to control door body when human body is classified as children and unlock, and door body is controlled when human body is classified as non-children
Keep locking.
Specifically, convolutional neural networks model includes three convolutional layers, two full articulamentums and a grader, grader
People's is classified as children and non-children corresponding to output human body.
The system further includes storage unit 25 and updating unit 26；Storage unit 25 is for storing convolutional neural networks mould
Type；Updating unit 26 is then used to be updated to be based on offline or online mode to convolutional neural networks model.
Structure, training and the application of specific convolutional neural networks model are described in detail in above-mentioned enabling control method,
It will not go into details herein.
The enabling control method and system that above-mentioned the application proposes, the body based on entire people are area-of-interest, are passed through
Convolutional neural networks, which automatically extract out the apparel characteristic of people, figure and features feature, height, fat or thin etc., can embody the characteristic value at age,
Judge the age, network output is two classifications of children or non-children, can it is accurate, efficiently classify to user's human body, have
Effect prevents children from arbitrarily opening door body, avoids potential danger.
It should be noted that it is limitation of the present invention that above description, which is not, the present invention is also not limited to the example above,
The variations, modifications, additions or substitutions that those skilled in the art are made in the essential scope of the present invention, are also answered
It belongs to the scope of protection of the present invention.
Claims (8)
1. enabling control method, which is characterized in that including：
Obtain human body region of interest area image；
Human body classification is carried out to the human body region of interest area image using convolutional neural networks model；
Door body is controlled when the human body is classified as children and keeps locking, and door body is controlled when the human body is classified as non-children and is opened
Lock.
2. enabling control method according to claim 1, which is characterized in that the convolutional neural networks model includes three
Convolutional layer, two full articulamentums and a grader, people's corresponding to grader output human body is classified as children and non-
It is virgin.
3. enabling control method according to claim 1, which is characterized in that the convolutional neural networks model can be based on from
Line or online mode update.
4. enabling control method according to claim 1, which is characterized in that the convolutional neural networks use stochastic gradient
Descent method is trained.
5. enabling control system, including door body, which is characterized in that further include camera unit, taxon and control unit；
The camera unit, for obtaining human body region of interest area image；
The taxon, for carrying out human body point to the human body region of interest area image using convolutional neural networks module
Class；
Described control unit keeps locking, in the human body point for controlling the door body when the human body is classified as children
Class controls the door body and unlocks when being non-children.
6. enabling control system according to claim 5, which is characterized in that the convolutional neural networks model includes three
Convolutional layer, two full articulamentums and a grader, people's corresponding to grader output human body is classified as children and non-
It is virgin.
7. enabling control system according to claim 5, which is characterized in that the system also includes storage units and update
Unit；
The storage unit, for storing the convolutional neural networks model；The updating unit, for to convolution god
Offline or online mode is based on through network model to be updated.
8. enabling control system according to claim 5, which is characterized in that the convolutional neural networks use stochastic gradient
Descent method is trained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810272557.9A CN108596171A (en) | 2018-03-29 | 2018-03-29 | Enabling control method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810272557.9A CN108596171A (en) | 2018-03-29 | 2018-03-29 | Enabling control method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596171A true CN108596171A (en) | 2018-09-28 |
Family
ID=63623926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810272557.9A Pending CN108596171A (en) | 2018-03-29 | 2018-03-29 | Enabling control method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596171A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109448176A (en) * | 2018-10-08 | 2019-03-08 | 厦门盈趣科技股份有限公司 | A kind of virgin lock starting method and device |
| CN109472894A (en) * | 2018-10-24 | 2019-03-15 | 常熟理工学院 | Distributed face recognition door lock system based on convolutional neural network |
| CN111487897A (en) * | 2019-01-25 | 2020-08-04 | 瑞萨电子株式会社 | Device control system and control method |
| CN112037410A (en) * | 2020-11-06 | 2020-12-04 | 上海兴容信息技术有限公司 | Control method and system of intelligent access control |
| CN112700576A (en) * | 2020-12-29 | 2021-04-23 | 成都启源西普科技有限公司 | Multi-modal recognition algorithm based on images and characters |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101363722A (en) * | 2008-09-25 | 2009-02-11 | 广州广电运通金融电子股份有限公司 | Height measurement method and measurement device thereof |
| US20100007726A1 (en) * | 2006-10-19 | 2010-01-14 | Koninklijke Philips Electronics N.V. | Method and apparatus for classifying a person |
| CN104504376A (en) * | 2014-12-22 | 2015-04-08 | 厦门美图之家科技有限公司 | Age classification method and system for face images |
| CN104992142A (en) * | 2015-06-03 | 2015-10-21 | 江苏大学 | Pedestrian recognition method based on combination of depth learning and property learning |
| CN105718873A (en) * | 2016-01-18 | 2016-06-29 | 北京联合大学 | People stream analysis method based on binocular vision |
| CN106337263A (en) * | 2016-11-14 | 2017-01-18 | Tcl集团股份有限公司 | A Washing Machine Control Method and Device |
| CN107045623A (en) * | 2016-12-30 | 2017-08-15 | 厦门瑞为信息技术有限公司 | A kind of method of the indoor dangerous situation alarm based on human body attitude trace analysis |
| CN107153871A (en) * | 2017-05-09 | 2017-09-12 | 浙江农林大学 | Fall detection method based on convolutional neural networks and mobile phone sensor data |
| CN107229893A (en) * | 2016-03-24 | 2017-10-03 | 杭州海康威视数字技术股份有限公司 | It whether there is the method and device of children in a kind of copilot room for detecting vehicle |
| CN107742107A (en) * | 2017-10-20 | 2018-02-27 | 北京达佳互联信息技术有限公司 | Facial image sorting technique, device and server |
-
2018
- 2018-03-29 CN CN201810272557.9A patent/CN108596171A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100007726A1 (en) * | 2006-10-19 | 2010-01-14 | Koninklijke Philips Electronics N.V. | Method and apparatus for classifying a person |
| CN101363722A (en) * | 2008-09-25 | 2009-02-11 | 广州广电运通金融电子股份有限公司 | Height measurement method and measurement device thereof |
| CN104504376A (en) * | 2014-12-22 | 2015-04-08 | 厦门美图之家科技有限公司 | Age classification method and system for face images |
| CN104992142A (en) * | 2015-06-03 | 2015-10-21 | 江苏大学 | Pedestrian recognition method based on combination of depth learning and property learning |
| CN105718873A (en) * | 2016-01-18 | 2016-06-29 | 北京联合大学 | People stream analysis method based on binocular vision |
| CN107229893A (en) * | 2016-03-24 | 2017-10-03 | 杭州海康威视数字技术股份有限公司 | It whether there is the method and device of children in a kind of copilot room for detecting vehicle |
| CN106337263A (en) * | 2016-11-14 | 2017-01-18 | Tcl集团股份有限公司 | A Washing Machine Control Method and Device |
| CN107045623A (en) * | 2016-12-30 | 2017-08-15 | 厦门瑞为信息技术有限公司 | A kind of method of the indoor dangerous situation alarm based on human body attitude trace analysis |
| CN107153871A (en) * | 2017-05-09 | 2017-09-12 | 浙江农林大学 | Fall detection method based on convolutional neural networks and mobile phone sensor data |
| CN107742107A (en) * | 2017-10-20 | 2018-02-27 | 北京达佳互联信息技术有限公司 | Facial image sorting technique, device and server |
Non-Patent Citations (3)
| Title |
|---|
| 俞一云等: "基于卷积神经网络的ADHD的判别分析", 《微型机与应用》 * |
| 李超琪等: "基于深度卷积神经网络的人脸年龄分类", 《智能计算机与应用》 * |
| 罗鸿斌: "行人再识别技术及其应用", 《电脑知识与技术》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109448176A (en) * | 2018-10-08 | 2019-03-08 | 厦门盈趣科技股份有限公司 | A kind of virgin lock starting method and device |
| CN109472894A (en) * | 2018-10-24 | 2019-03-15 | 常熟理工学院 | Distributed face recognition door lock system based on convolutional neural network |
| CN111487897A (en) * | 2019-01-25 | 2020-08-04 | 瑞萨电子株式会社 | Device control system and control method |
| CN112037410A (en) * | 2020-11-06 | 2020-12-04 | 上海兴容信息技术有限公司 | Control method and system of intelligent access control |
| CN112700576A (en) * | 2020-12-29 | 2021-04-23 | 成都启源西普科技有限公司 | Multi-modal recognition algorithm based on images and characters |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596042A (en) | Enabling control method and system | |
| CN108596171A (en) | Enabling control method and system | |
| Casey et al. | You might be a robot | |
| Mohanty | The epistemic status of cultural identity | |
| Chen et al. | Learning to explain: An information-theoretic perspective on model interpretation | |
| CN109492560A (en) | Facial image Feature fusion, device and storage medium based on time scale | |
| CN109670446A (en) | Anomaly detection method based on linear dynamic system and depth network | |
| CN111814706A (en) | Face recognition and attribute classification method based on multi-task convolutional neural network | |
| CN107220635A (en) | Human face in-vivo detection method based on many fraud modes | |
| CN103678965B (en) | A kind of method and device of protection mobile device safety | |
| CN107766785A (en) | A kind of face recognition method | |
| CN107292267A (en) | Photo fraud convolutional neural networks training method and human face in-vivo detection method | |
| Yang et al. | Moving object detection for dynamic background scenes based on spatiotemporal model | |
| CN108763542A (en) | A kind of Text Intelligence sorting technique, device and computer equipment based on combination learning | |
| Levin | Imaginability, possibility, and the puzzle of imaginative resistance | |
| US20230259560A1 (en) | Apparatus, method and computer program for creating digital memories for a particular person | |
| CN110288731A (en) | Unlocking method and device and electronic door lock | |
| CN109242089A (en) | Progressive supervised deep learning neural network training method, system, medium and device | |
| CN112002075A (en) | An information processing method and device for improving the security of a locker | |
| Laurence et al. | What happens to our representation of identity as familiar faces age? Evidence from priming and identity aftereffects | |
| CN110113423A (en) | A method of message is used based on internet push intelligent door lock | |
| Versfeld | AllPay and no work: spheres of belonging under duress | |
| Humbach | Neuroscience, Justice and the Mental Causation Fallacy | |
| Baig et al. | Bio-inspired probabilistic model for crowd emotion detection | |
| Hu et al. | Research and Prospect of Image Recognition Based on Convolutional Neural Network |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |