CN106815563A - A kind of crowd's quantitative forecasting technique based on human body apparent structure - Google Patents
A kind of crowd's quantitative forecasting technique based on human body apparent structure Download PDFInfo
- Publication number
- CN106815563A CN106815563A CN201611225785.8A CN201611225785A CN106815563A CN 106815563 A CN106815563 A CN 106815563A CN 201611225785 A CN201611225785 A CN 201611225785A CN 106815563 A CN106815563 A CN 106815563A
- Authority
- CN
- China
- Prior art keywords
- crowd
- pedestrian
- scene
- image
- quantitative
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Abstract
The invention discloses a kind of crowd's quantitative forecasting technique based on human body apparent structure, for predicting the crowd's quantity in given scenario image.Specifically include following steps:Obtain for training the monitoring image data set of crowd's quantitative forecast model, and define algorithm target;Apparent semantic structure to pedestrian's body in monitoring image data set is modeled, and Density Distribution and body shape to pedestrian carries out joint modeling;Modeling result in step S2 sets up the forecast model of crowd's quantity;The crowd's quantity in scene image is predicted using the forecast model.Crowd quantitative forecast of the present invention suitable for real video monitoring scene, has preferably effect and robustness in face of all kinds of complex situations.
Description
Technical field
The invention belongs to computer vision field, a kind of particularly crowd's quantitative forecast based on human body apparent structure
Method.
Background technology
Since 20 end of the centurys, with the development of computer vision, intelligent Video Surveillance Technology is widely paid close attention to and is ground
Study carefully.It is the important and challenging task of one of which that crowd counts, during its target is Accurate Prediction Dense crowd image
Pedestrian's quantity.Three key factors of crowd's counting load are pedestrian, head and their context mechanism.When us
In the quantity of the crowd of statistics, we can accurately be judged the mankind by the use of the semantic structure of pedestrian's body different piece as clue
Everyone position.Therefore, prediction crowd's quantity needs to be analyzed the semantic structure of pedestrian's body exactly.
Existing people counting method generally comprises following three types:1st, the crowd based on pedestrian detector counts.This kind of side
Method matches each pedestrian in image using various pedestrian detectors;2nd, counted based on the global crowd for returning.This kind of method
Mapping between main modeling crowd's image and crowd's quantity;3rd, the crowd based on density estimation counts.This kind of Method Modeling people
The Density Distribution of group, then crowd's quantity is predicted by Density Distribution.Existing method is whole as one using the whole body of pedestrian
Volume modeling, or only model the head of pedestrian.They have ignored the semantic structure information of abundant pedestrian's body part, using this
A little structural informations can improve the performance of crowd's counting algorithm.
The content of the invention
To solve the above problems, it is an object of the invention to provide a kind of crowd's quantitative forecast based on human body apparent structure
Method, for predicting the crowd's quantity in given scenario image.This method is based on physical appearance of the deep neural network to pedestrian
Structure and Density Distribution information carry out semantic modeling, and predict accurate crowd's quantity according to modeling result, can preferably fit
Answer the complex situations in real video monitoring scene.
To achieve the above object, the technical scheme is that:
A kind of crowd's quantitative forecasting technique based on human body apparent structure, comprises the following steps:
S1, obtain for training the monitoring image data set of crowd's quantitative forecast model, and define algorithm target;
S2, the apparent semantic structure to pedestrian's body in monitoring image data set are modeled, and to the density point of pedestrian
Cloth and body shape carry out joint modeling;
S3, the modeling result in step S2 set up the forecast model of crowd's quantity;
S4, use the forecast model predict scene image in crowd's quantity.
Further, in step S1, the described monitoring image data set for training crowd's quantitative forecast model, including
Scene imageThe pedestrian head position P of artificial marktrainWith scene depth figure
Defining algorithm target is:Predict a width scene imageIn pedestrian's quantity
Further, in step S2, the apparent semantic structure of pedestrian's body is modeled and is specifically included:
S21, according to all pedestrian head position P in monitoring image data settrainAnd its respective scene depth valuePosition and the size of each pedestrian image bounding box are determined, with this from scene graph image setMiddle sanction
Cut to obtain pedestrian image Itrain;
S22, by pedestrian image ItrainInput single row people semanteme segmenting system carries out semantic segmentation respectively;
S23, to every width scene imageThe segmentation result of wherein all pedestrians is pressed into full size and position recovering, is obtained
Scene imageCrowd's semantic structure figure Reflection scene imageIn all pedestrians body part semantic structure letter
Breath.
Further, in step S2, the joint modeling of Density Distribution and body shape to pedestrian is specifically included:
S24, to scene imageIn pedestrian Density Distribution and body shape carry out joint modeling, obtain structuring people
Population density figure
Wherein, p isOn location of pixels,It is shape of the two-dimensional Gaussian kernel to the approximate number of people,It is two-dimentional high
This core is used to the shape of the approximate person,WithIt is respectively the center of i-th number of people and the person,Take from
Ptrain,ByWith scene depth valueEstimation draws, σhAnd σbIt is respectivelyWithVariance, they respectively byWithEstimation is obtained,By crowd's semantic structure figureBinaryzation is obtained,It is the pedestrian's number in scene
Amount, Z is that normalization coefficient makes each pedestrian existOn density and be 1, structuring crowd density figureReflection scene imageIn all pedestrians Density Distribution and body shape information.
Further, in step S3, the forecast model for setting up crowd's quantity is specifically included:
S31, depth convolutional neural networks are set up, the input of neutral net is a width scene imageIt is output as correspondence
Crowd's semantic structure figureStructuring crowd density figureAndIn pedestrian's quantitySo as to the structure of neutral net
Mapping can be expressed as
S32, sub- mappingUsing soft maximum (Softmax) loss function, it is expressed as
WhereinIt is one of output of neutral net,RepresentMiddle location of pixels (h, w) and the value of passage i,Generated by step S23 methods describeds,RepresentThe value of middle location of pixels (h, w);
S33, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net,Generated by step S24 methods describeds;
S34, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net,It is crowd's quantity of artificial mark;
S35, the loss function of whole neutral net are
L=Lc+λdLd+λbLbFormula (5)
Whole neutral net is trained under loss function L using stochastic gradient descent and back-propagation algorithm.
Further, in step S4, the crowd's quantity in prediction scene image includes:By scene image to be predictedIt is defeated
Enter the neutral net for training, crowd's quantity of its outputThe as result of crowd's quantitative forecast.
Crowd's quantitative forecasting technique based on human body apparent structure of the invention, compared to existing crowd quantitative forecast side
Method, has the advantages that:
First, crowd's quantitative forecasting technique of the invention has excavated the semantic attribute of crowd's enumeration problem, defines and models
Three key factors of this problem:Body, head and their context mechanism.This kind is assumed in more adaptation actual scene
Complex situations.
Secondly, crowd's quantitative forecasting technique of the invention sets up crowd's quantitative forecast mould based on depth convolutional neural networks
Type.Depth convolutional neural networks can preferably express visual signature, in addition, Visual Feature Retrieval Process, pedestrian's semantic modeling and people
Group's quantity is returned and is unified in same framework, improves the final effect of method.
Crowd's quantitative forecasting technique based on human body apparent structure of the invention, has in intelligent video monitoring analysis system
There is good application value, efficiency and the degree of accuracy of crowd's quantitative forecast can be effectively improved.For example, in the application of public safety
In scene, crowd's quantitative forecasting technique of the invention can quickly and correctly predicting monitoring camera shooting area pedestrian's number
Amount, is that the day-to-day operation of public place and emergency processing provide decision-making foundation.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the crowd's quantitative forecasting technique based on human body apparent structure of the invention.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Conversely, the present invention covers any replacement done in spirit and scope of the invention being defined by the claims, repaiies
Change, equivalent method and scheme.Further, in order that the public has a better understanding to the present invention, below to of the invention thin
It is detailed to describe some specific detail sections in section description.Part without these details for a person skilled in the art
Description can also completely understand the present invention.
With reference to Fig. 1, in the preferred embodiment, the crowd's quantitative forecasting technique based on human body apparent structure, bag
Include following steps:
First, the monitoring image data set for training crowd's quantitative forecast model is obtained.Wherein, for training crowd's number
Measure the monitoring image data set of forecast model, including scene imageThe pedestrian head position P of artificial marktrainAnd field
Scape depth map
Defining algorithm target is:Predict a width scene imageIn pedestrian's quantity
Secondly, the Density Distribution and body shape to pedestrian in the monitoring image data set of acquisition carry out joint modeling.Tool
Body, it comprises the following steps:
The first step, according to all pedestrian head position P in monitoring image data settrainAnd its respective scene depth valuePosition and the size of each pedestrian image bounding box are determined, with this from scene graph image setMiddle sanction
Cut to obtain pedestrian image Itrain;
Second step, by pedestrian image ItrainInput single row people semanteme segmenting system carries out semantic segmentation respectively;
3rd step, to every width scene imageThe segmentation result of wherein all pedestrians is pressed into full size and position recovering, is obtained
To scene imageCrowd's semantic structure figure Reflection scene imageIn all pedestrians body part semantic structure
Information.
Next, carrying out joint modeling to the Density Distribution and body shape of pedestrian.To scene imageIn pedestrian
Density Distribution and body shape carry out joint modeling, obtain structuring crowd density figure
Wherein, p isOn location of pixels,It is shape of the two-dimensional Gaussian kernel to the approximate number of people,It is two-dimentional high
This core is used to the shape of the approximate person.WithIt is respectively the center of i-th number of people and the person,Take from
Ptrain,ByWith scene depth valueEstimation draws.σhAnd σbIt is respectivelyWithVariance, they respectively byWithEstimation is obtained.By crowd's semantic structure figureBinaryzation is obtained.It is the pedestrian's number in scene
Amount, Z is that normalization coefficient makes each pedestrian existOn density and be 1.Structuring crowd density figureReflection scene imageIn all pedestrians Density Distribution and body shape information.
Afterwards, the forecast model of crowd's quantity is set up.Specifically include:
The first step, sets up depth convolutional neural networks, and the input of neutral net is a width scene imageIt is output as correspondenceCrowd's semantic structure figureStructuring crowd density figureAndIn pedestrian's quantitySo as to the knot of neutral net
Structure can be expressed as mapping
Second step, sub- mappingUsing soft maximum (Softmax) loss function, it is expressed as
WhereinIt is one of output of neutral net,RepresentMiddle location of pixels (h, w) and the value of passage i,RepresentThe value of middle location of pixels (h, w);
3rd step, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net,Generated by formula (1) methods described.
4th step, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net,It is crowd's quantity of artificial mark.
5th step, the loss function of whole neutral net is
L=Lc+λdLd+λbLbFormula (5)
Whole neutral net is trained under loss function L using stochastic gradient descent and back-propagation algorithm.
Finally, the crowd's quantity in scene image to be predicted is predicted using the model set up.Specifically include:Will be pre-
The scene image of surveyThe neutral net that input is trained, crowd's quantity of its outputThe as result of crowd's quantitative forecast.
In above-described embodiment, crowd's quantitative forecasting technique of the invention is first by the physical appearance structure and density of pedestrian point
Cloth information modeling is two kinds of semantic scene models.On this basis, former problem is converted into multi-task learning problem, and based on depth
Degree neural network crowd's quantitative forecast model.Finally, new field is predicted using the crowd's quantitative forecast model for training
Accurate pedestrian's quantity in scape image.
By above technical scheme, the embodiment of the present invention has developed one kind and has been applied to video monitoring based on depth learning technology
Crowd's quantitative forecast algorithm of scene.The present invention can effectively model the body semantic structure information and Density Distribution of pedestrian simultaneously
Information, so as to predict accurate crowd's quantity.
Presently preferred embodiments of the present invention is the foregoing is only, is not intended to limit the invention, it is all in essence of the invention
Any modification, equivalent and improvement made within god and principle etc., should be included within the scope of the present invention.
Claims (6)
1. a kind of crowd's quantitative forecasting technique based on human body apparent structure, it is characterised in that comprise the following steps:
S1, obtain for training the monitoring image data set of crowd's quantitative forecast model, and define algorithm target;
S2, the apparent semantic structure to pedestrian's body in monitoring image data set are modeled, and Density Distribution to pedestrian and
Body shape carries out joint modeling;
S3, the modeling result in step S2 set up the forecast model of crowd's quantity;
S4, use the forecast model predict scene image in crowd's quantity.
2. crowd's quantitative forecasting technique of human body apparent structure is based on as claimed in claim 1, it is characterised in that step S1
In, the described monitoring image data set for training crowd's quantitative forecast model, including scene imageArtificial mark
Pedestrian head position PtrainWith scene depth figure
Defining algorithm target is:Predict a width scene imageIn pedestrian's quantity C.
3. crowd's quantitative forecasting technique of human body apparent structure is based on as claimed in claim 2, it is characterised in that step S2
In, the apparent semantic structure of pedestrian's body is modeled and is specifically included:
S21, according to all pedestrian head position P in monitoring image data settrainAnd its respective scene depth valuePosition and the size of each pedestrian image bounding box are determined, with this from scene graph image setMiddle sanction
Cut to obtain pedestrian image Itrain;
S22, by pedestrian image ItrainInput single row people semanteme segmenting system carries out semantic segmentation respectively;
S23, to every width scene imageThe segmentation result of wherein all pedestrians is pressed into full size and position recovering, scene graph is obtained
PictureCrowd's semantic structure figure Reflection scene imageIn all pedestrians body part semantic structure information.
4. crowd's quantitative forecasting technique of human body apparent structure is based on as claimed in claim 3, it is characterised in that step S2
In, joint modeling is carried out to the Density Distribution and body shape of pedestrian and is specifically included:
S24, to scene imageIn pedestrian Density Distribution and body shape carry out joint modeling, obtain structuring crowd close
Degree figure
Wherein, p isOn location of pixels,It is shape of the two-dimensional Gaussian kernel to the approximate number of people,It is two-dimensional Gaussian kernel
To the shape of the approximate person,WithIt is respectively the center of i-th number of people and the person,Take from Ptrain,
ByWith scene depth valueEstimation draws, σhAnd σbIt is respectivelyWithVariance, respectively byWithEstimation is obtained,By crowd's semantic structure figureBinaryzation is obtained, and C is the pedestrian's quantity in scene, and Z is normalization
Coefficient makes each pedestrian existOn density and be 1, structuring crowd density figureReflection scene imageIn all pedestrians
Density Distribution and body shape information.
5. crowd's quantitative forecasting technique of human body apparent structure is based on as claimed in claim 4, it is characterised in that step S3
In, the forecast model for setting up crowd's quantity is specifically included:
S31, depth convolutional neural networks are set up, the input of neutral net is a width scene imageIt is output as correspondenceCrowd
Semantic structure figureStructuring crowd density figureAndIn pedestrian's quantitySo as to the structure of neutral net can be with table
It is shown as mapping
S32, sub- mappingUsing soft maximum (Softmax) loss function, it is expressed as
WhereinIt is one of output of neutral net,RepresentMiddle location of pixels (h, w) and the value of passage i,By
Step S23 methods describeds are generated,RepresentThe value of middle location of pixels (h, w);
S33, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net,Generated by step S24 methods describeds;
S34, sub- mappingUsing Euclid's loss function, it is expressed as
WhereinIt is one of output of neutral net, C is crowd's quantity of artificial mark;
S35, the loss function of whole neutral net are
L=Lc+λdLd+λbLbFormula (5)
Whole neutral net is trained under loss function L using stochastic gradient descent and back-propagation algorithm.
6. crowd's quantitative forecasting technique of human body apparent structure is based on as claimed in claim 5, it is characterised in that step S4
In, the crowd's quantity in prediction scene image includes:By scene image to be predictedThe neutral net that input is trained, its is defeated
The result of the crowd's quantity C as crowd's quantitative forecasts for going out.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611225785.8A CN106815563B (en) | 2016-12-27 | 2016-12-27 | Human body apparent structure-based crowd quantity prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611225785.8A CN106815563B (en) | 2016-12-27 | 2016-12-27 | Human body apparent structure-based crowd quantity prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106815563A true CN106815563A (en) | 2017-06-09 |
CN106815563B CN106815563B (en) | 2020-06-02 |
Family
ID=59110304
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611225785.8A Active CN106815563B (en) | 2016-12-27 | 2016-12-27 | Human body apparent structure-based crowd quantity prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106815563B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107622244A (en) * | 2017-09-25 | 2018-01-23 | 华中科技大学 | A kind of indoor scene based on depth map becomes more meticulous analytic method |
CN109508583A (en) * | 2017-09-15 | 2019-03-22 | 杭州海康威视数字技术股份有限公司 | A kind of acquisition methods and device of distribution trend |
CN109961060A (en) * | 2019-04-11 | 2019-07-02 | 北京百度网讯科技有限公司 | Method and apparatus for generating crowd density information |
CN110505440A (en) * | 2018-05-18 | 2019-11-26 | 杭州海康威视数字技术股份有限公司 | A kind of area monitoring method and device |
CN112026686A (en) * | 2019-06-04 | 2020-12-04 | 上海汽车集团股份有限公司 | Method and device for automatically adjusting position of vehicle seat |
CN115083112A (en) * | 2022-08-22 | 2022-09-20 | 枫树谷(成都)科技有限责任公司 | Intelligent early warning emergency management system and deployment method thereof |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101976353A (en) * | 2010-10-28 | 2011-02-16 | 北京智安邦科技有限公司 | Statistical method and device of low density crowd |
CN102063613A (en) * | 2010-12-28 | 2011-05-18 | 北京智安邦科技有限公司 | People counting method and device based on head recognition |
CN103020606A (en) * | 2012-12-27 | 2013-04-03 | 北京大学 | Pedestrian detection method based on spatio-temporal context information |
CN103093211A (en) * | 2013-01-27 | 2013-05-08 | 西安电子科技大学 | Human motion tracking method based on deep nuclear information image feature |
CN103646257A (en) * | 2013-12-30 | 2014-03-19 | 中国科学院自动化研究所 | Video monitoring image-based pedestrian detecting and counting method |
US20150285639A1 (en) * | 2014-04-04 | 2015-10-08 | Umm-Al-Qura University | Method and system for crowd sensing to be used for automatic semantic identification |
CN105184260A (en) * | 2015-09-10 | 2015-12-23 | 北京大学 | Image characteristic extraction method, pedestrian detection method and device |
CN106066993A (en) * | 2016-05-23 | 2016-11-02 | 上海交通大学 | A kind of crowd's semantic segmentation method and system |
-
2016
- 2016-12-27 CN CN201611225785.8A patent/CN106815563B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101976353A (en) * | 2010-10-28 | 2011-02-16 | 北京智安邦科技有限公司 | Statistical method and device of low density crowd |
CN102063613A (en) * | 2010-12-28 | 2011-05-18 | 北京智安邦科技有限公司 | People counting method and device based on head recognition |
CN103020606A (en) * | 2012-12-27 | 2013-04-03 | 北京大学 | Pedestrian detection method based on spatio-temporal context information |
CN103093211A (en) * | 2013-01-27 | 2013-05-08 | 西安电子科技大学 | Human motion tracking method based on deep nuclear information image feature |
CN103646257A (en) * | 2013-12-30 | 2014-03-19 | 中国科学院自动化研究所 | Video monitoring image-based pedestrian detecting and counting method |
US20150285639A1 (en) * | 2014-04-04 | 2015-10-08 | Umm-Al-Qura University | Method and system for crowd sensing to be used for automatic semantic identification |
CN105184260A (en) * | 2015-09-10 | 2015-12-23 | 北京大学 | Image characteristic extraction method, pedestrian detection method and device |
CN106066993A (en) * | 2016-05-23 | 2016-11-02 | 上海交通大学 | A kind of crowd's semantic segmentation method and system |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109508583A (en) * | 2017-09-15 | 2019-03-22 | 杭州海康威视数字技术股份有限公司 | A kind of acquisition methods and device of distribution trend |
CN109508583B (en) * | 2017-09-15 | 2020-11-06 | 杭州海康威视数字技术股份有限公司 | Method and device for acquiring crowd distribution characteristics |
CN107622244A (en) * | 2017-09-25 | 2018-01-23 | 华中科技大学 | A kind of indoor scene based on depth map becomes more meticulous analytic method |
CN107622244B (en) * | 2017-09-25 | 2020-08-28 | 华中科技大学 | Indoor scene fine analysis method based on depth map |
CN110505440A (en) * | 2018-05-18 | 2019-11-26 | 杭州海康威视数字技术股份有限公司 | A kind of area monitoring method and device |
CN109961060A (en) * | 2019-04-11 | 2019-07-02 | 北京百度网讯科技有限公司 | Method and apparatus for generating crowd density information |
CN109961060B (en) * | 2019-04-11 | 2021-04-30 | 北京百度网讯科技有限公司 | Method and apparatus for generating crowd density information |
CN112026686A (en) * | 2019-06-04 | 2020-12-04 | 上海汽车集团股份有限公司 | Method and device for automatically adjusting position of vehicle seat |
CN115083112A (en) * | 2022-08-22 | 2022-09-20 | 枫树谷(成都)科技有限责任公司 | Intelligent early warning emergency management system and deployment method thereof |
Also Published As
Publication number | Publication date |
---|---|
CN106815563B (en) | 2020-06-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111611878B (en) | Method for crowd counting and future people flow prediction based on video image | |
CN110502965B (en) | Construction safety helmet wearing monitoring method based on computer vision human body posture estimation | |
CN110781838B (en) | Multi-mode track prediction method for pedestrians in complex scene | |
CN106815563A (en) | A kind of crowd's quantitative forecasting technique based on human body apparent structure | |
CN107967451B (en) | Method for counting crowd of still image | |
CN110147743A (en) | Real-time online pedestrian analysis and number system and method under a kind of complex scene | |
CN108830145B (en) | People counting method based on deep neural network and storage medium | |
CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
CN108921822A (en) | Image object method of counting based on convolutional neural networks | |
CN110210551A (en) | A kind of visual target tracking method based on adaptive main body sensitivity | |
CN110334705A (en) | A kind of Language Identification of the scene text image of the global and local information of combination | |
CN103577875B (en) | A kind of area of computer aided CAD demographic method based on FAST | |
CN105005760B (en) | A kind of recognition methods again of the pedestrian based on Finite mixture model | |
CN105550678A (en) | Human body motion feature extraction method based on global remarkable edge area | |
CN109978918A (en) | A kind of trajectory track method, apparatus and storage medium | |
CN111783589B (en) | Complex scene crowd counting method based on scene classification and multi-scale feature fusion | |
CN107122736A (en) | A kind of human body based on deep learning is towards Forecasting Methodology and device | |
CN107301376B (en) | Pedestrian detection method based on deep learning multi-layer stimulation | |
CN111191667A (en) | Crowd counting method for generating confrontation network based on multiple scales | |
CN110047081A (en) | Example dividing method, device, equipment and the medium of chest x-ray image | |
CN116258608B (en) | Water conservancy real-time monitoring information management system integrating GIS and BIM three-dimensional technology | |
CN114519302A (en) | Road traffic situation simulation method based on digital twin | |
CN102509119B (en) | Method for processing image scene hierarchy and object occlusion based on classifier | |
CN109614896A (en) | A method of the video content semantic understanding based on recursive convolution neural network | |
CN109063549A (en) | High-resolution based on deep neural network is taken photo by plane video moving object detection method |
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 |