CN107862261A - Image people counting method based on multiple dimensioned convolutional neural networks - Google Patents

Image people counting method based on multiple dimensioned convolutional neural networks Download PDF

Info

Publication number
CN107862261A
CN107862261A CN201711014291.XA CN201711014291A CN107862261A CN 107862261 A CN107862261 A CN 107862261A CN 201711014291 A CN201711014291 A CN 201711014291A CN 107862261 A CN107862261 A CN 107862261A
Authority
CN
China
Prior art keywords
mrow
neural networks
convolutional neural
density map
image
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
Application number
CN201711014291.XA
Other languages
Chinese (zh)
Inventor
周圆
杨建兴
李成浩
杜晓婷
毛爱玲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin University filed Critical Tianjin University
Priority to CN201711014291.XA priority Critical patent/CN107862261A/en
Publication of CN107862261A publication Critical patent/CN107862261A/en
Pending legal-status Critical Current

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
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of image people counting method based on multiple dimensioned convolutional neural networks, step (1), the continuous density map label of generation, the image marked is converted into continuous estimation density map；Step (2), the accurate density map of prediction crowd is obtained using multiple dimensioned convolutional neural networks, after for convolutional neural networks, one initial parameter is set, the loss L (θ) of input picture is calculated according to the density map of reality, then the parameter of whole network is updated in Optimized Iterative each time, until penalty values converge to a less value.Compared with prior art, the present invention solves crowd's enormousness in single image and changed, on the basis of single branch convolutional neural networks, the feature of different levels network has been merged before predicted density figure is generated, the feature that different depth corresponds to different scale is extracted, greatly improves the precision of predicted density figure；The problems such as solving the dimensional variation in crowd's image and blocking.

Description

Image people counting method based on multiple dimensioned convolutional neural networks
Technical field
The present invention relates to crowd's image analysis technology field, specifically a kind of crowd based on multiple dimensioned convolutional neural networks Counting algorithm.
Background technology
It is a kind of intelligent monitoring application of the quantity of the density map calculating people by predicting crowd's image that crowd, which counts,.With The exponential increase of world population, quick urbanization promote many large-scale activities, such as sports match, Gong Zhongyou OK, the problems such as congested in traffic, causes large-scale crowd massing.So in order to preferably manage crowd and personal safety, Ren Qunhang It is significant for parser.
With the continuous popularization of deep learning algorithm, crowd's counting algorithm contrast traditional algorithm based on convolutional neural networks Substantially increase accuracy of detection.Algorithm based on convolutional neural networks is broadly divided into two kinds：A kind of is the algorithm based on recurrence, separately A kind of is the algorithm based on density map.The former is to be used as label, training convolutional nerve net with corresponding number by the use of crowd's image Network study maps to one from crowd's image to the nonlinear function of crowd's quantity, and the output of network is the number of crowd.The latter It is by the use of crowd's image and corresponding density map as label, goes training convolutional neural networks generation corresponding with input crowd's image Density map, different from the method for recurrence, the network of the algorithm based on density map is using density map as output, according to prediction Density map goes calculating crowd's quantity.But because crowd's image is deposited in monitoring camera and high-altitude shooting, shooting angle mostly In great changes, the image taken there is very big change in the size of people and yardstick.The multiple row convolution god that Zhang et al. is proposed Through network in network complexity it is very high, network parameter is very big, and three row networks need pre-training again by multiple row network output characteristic Merged, it is impossible to while hold the multi-scale information of single image.
The content of the invention
The present invention seeks to extract the feature of different depth for Tilly convolutional neural networks, different scale feature is melted Close, it is proposed that a kind of crowd density detection method based on multiple dimensioned convolutional neural networks, it is close by being predicted from crowd's image Degree figure calculates total number.
A kind of image people counting method based on multiple dimensioned convolutional neural networks of the present invention, this method include following step Suddenly：
Step 1, the continuous density map label of generation, specifically include following processing：
By density map corresponding to the good number of people Coordinate generation of handmarking, there is the graphical representation of N number of people's labeling head for such as Minor function：
In formula, δ (x-xi) it is delta function；xiRepresent the position where people's leader note point；
The image marked is converted into continuous density map, expression formula is as follows：
F (x)=H (x)*
Step 2, the accurate density map of prediction crowd is obtained using multiple dimensioned convolutional neural networks, specifically include following place Reason：
Multiple dimensioned convolutional neural networks obtain three convolutional layers by the connection in pond of the convolution of convolution-pond-again-again, from First three convolutional layer extracts the wild feature of different feeling, and these features are merged in a manner of cascading merging, then passes through Density map corresponding to two convolutional layer outputs；
The loss function L (θ) of the multiple dimensioned convolutional neural networks is calculated, expression formula is as follows：
Wherein, N be input convolutional neural networks amount of images, xiFor the i-th width input picture of convolutional neural networks, M (xi) represent the i-th width input picture standard density figure matrix；
After one initial parameter is set for convolutional neural networks, the loss L of input picture is calculated according to the density map of reality (θ), the parameter of whole network is then updated in Optimized Iterative each time, until penalty values converge to a less value.
Compared with prior art, the image people counting method of the invention based on multiple dimensioned convolutional neural networks have with Lower effect：
1st, single-row convolutional neural networks can be utilized in the case of compared with low parameter, with reference to the feature of different depth, detection The pedestrian of different scale into crowd's image；
2nd, solve crowd's enormousness change in single image, on the basis of single branch convolutional neural networks, generating The feature of different levels network has been merged before predicted density figure, has extracted the feature that different depth corresponds to different scale, greatly Improve the precision of predicted density figure；
3rd, the problems such as solving the dimensional variation in crowd's image and blocking.
Brief description of the drawings
Fig. 1 is the image people counting method overall flow schematic diagram based on multiple dimensioned convolutional neural networks of the present invention；
Fig. 2 is multiple dimensioned convolutional neural networks structure chart；
Fig. 3 is experimental result picture；It is crowd's image to scheme (a), and figure (b) is corresponding density map.
Embodiment
Embodiments of the present invention are described in further detail below in conjunction with accompanying drawing.
As shown in figure 1, a kind of crowd density detection method based on multiple dimensioned convolutional neural networks of the present invention, will be single-row Convolutional neural networks are merged in the feature of different depth, are comprised the following steps that：
Step 1, the continuous density map label of generation, the image marked is converted into continuous estimation density map, specifically Including following processing：
By density map corresponding to the good number of people Coordinate generation of handmarking, there is the graphical representation of N number of people's labeling head for such as Minor function：
In formula, δ (x-xi) it is delta function；xiRepresent the position where people's leader note point；
Estimate that density map F (x) expression formula is as follows：
F (x)=H (x)*
；
Step 2, the accurate density map of prediction crowd is obtained using multiple dimensioned convolutional neural networks：Multiple dimensioned convolutional Neural net Network obtains three convolutional layers by the connection in pond of the convolution of convolution-pond-again-again, and the not same feeling is extracted from first three convolutional layer By the feature of open country, those features are extracted multi-level feature by the convolutional layer of three different depths and formed, with adding for network Deep, the receptive field of higher convolutional layer also can be bigger, can be obtained in the feature that the convolutional layer of low level extracts more small The detailed information of object, what is obtained in high-level convolutional layer is advanced semantic feature, by these features to cascade merging Mode is merged, i.e. the superposition of characteristic pattern, then by density map corresponding to two convolutional layer outputs.The loss function of the network It is estimation density map F (xi；θ) and actual density figure M (xi) between Euclidean distance L (θ), expression is as follows：
Wherein, N be input convolutional neural networks amount of images, xiFor the i-th width input picture of convolutional neural networks, M (xi) represent the i-th width input picture accurate density map matrix；
After one initial parameter is set for convolutional neural networks, the damage of input picture is calculated according to the accurate density map of reality L (θ) is lost, the parameter of whole network is then updated in Optimized Iterative each time, until penalty values converge to a less value.
Due to camera shooting angle, different degrees of perspective distortion often occurs for crowd's image, its total body surface It is now that the area that the pedestrian nearer apart from camera occupies in the picture is larger, the area that the pedestrian away from camera occupies in the picture It is smaller.In this step, using the pedestrian of different scale in multiple dimensioned convolutional neural networks Monitoring Population image.In convolutional Neural In network, the feature of different depth represents different grades of feature in network.What convolutional neural networks were extracted in low layer is figure The profile and shape facility of picture, receptive field is relatively small, with the intensification of the network number of plies, deep layer network extraction to be image High-level semantics features, the feature of different levels in network is overlapped fusion, combined well multiple dimensioned in crowd's image Feature, it is final to produce the crowd density figure that more calculates to a nicety.
Specific embodiment is described as follows：
The present invention needs to solve the problems, such as to be " to give the frame in crowd's image or video, then estimate the figure As the density and number of regional crowd amount to "：
Known input picture is expressed as to M × N matrix：x∈Rm×n, then the actual crowd corresponding to input picture x is close Degree is expressed as：
Wherein, N represents the number in image, and x represents the position of each pixel in image, xiIt is i-th of number of people in image In position, δ (x-xi) impulse function is represented, * represents convolution operation symbol, Gδ(x) Gaussian kernel that standard deviation is δ is represented.
The target of the embodiment is one mapping function by input picture x to crowd density figure of study：
F:x→F(x)≈M(x)
Wherein, F (x) is estimation crowd density figure.
In order to learn F, it is necessary to optimize following problem：
Wherein, F (x；It is θ) estimation crowd density figure, θ is parameter to be learned.In general, F is a complex nonlinear Function.
As shown in Fig. 2 by the present invention using learning the multiple dimensioned of the nonlinear function F from crowd image to density map Convolutional neural networks.Multiple dimensioned convolutional neural networks are to be merged the feature of different depth level.By single-row convolutional Neural By a convolution, pond, second layer characteristic pattern pass through pond of a convolution to the first layer characteristic pattern of network twice, by before Two layers of obtained feature links together with the characteristic pattern that third layer convolution obtains in " passage " dimension, forms total characteristic figure Merged feature maps, then obtain density map to the end by two convolutional layers again.
The loss function of above-mentioned multiple dimensioned convolutional neural networks be estimate Euclidean between density map and actual density figure away from From：
Update the parameter L (θ) of whole network in training process in Optimized Iterative each time using gradient descent method, until Penalty values converge to a less value.
The present invention compares in three common data sets with other method, including market data set MALL, UCSD With SHANGHAITECH data sets.The evaluation criterion of experimental result uses：
Mean absolute error (MAE)：
With mean square error (MSE)：
N is picture number, ziFor number of people number actual in the i-th width image,Pass through for the i-th width image provided by the invention The number of people number of network output) carry out the accuracy of measure algorithm.On the data set of MALL markets, the of the invention and technology of existing algorithm Contrast, (wherein MD-CNN is inventive algorithm) as shown in table 1：
Table 1
On UCSD data sets, the present invention is compared with the prior art, as shown in table 2：
Table 2
Method MAE MSE
Kernelridgeregression 2.16 7.45
Ridgeregression 2.25 7.82
Gaussianprocessregression 2.24 7.97
Cumulativeattributeregression 2.07 6.86
Zhangetal. 1.60 3.31
MCNN 1.07 1.35
MDCNN(ours) 1.16 1.75
It is as shown in table 3 with the comparison of other existing algorithms on SHANGHAITECH part_B data sets：
Table 3
Method MAE MSE
LBP+RR 59.1 87.1
Zhangetal. 32 49.8
MCNN 26.4 41.3
MDCNN(ours) 22.3 39.45

Claims (1)

1. a kind of image people counting method based on multiple dimensioned convolutional neural networks, it is characterised in that this method includes following Step：
Step (1), the continuous density map label of generation, the image marked is converted into continuous estimation density map, specific bag Include following processing：
By density map corresponding to the good number of people Coordinate generation of handmarking, the graphical representation with N number of people's labeling head is following letter Number：
<mrow> <mi>H</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <mi>&amp;delta;</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>-</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow>
In formula, δ (x-xi) it is delta function；xiRepresent the position where people's leader note point；
Estimate that density map F (x) expression formula is as follows：
F (x)=H (x)*
；
Step (2), the accurate density map of prediction crowd is obtained using multiple dimensioned convolutional neural networks, specifically include following processing：
Multiple dimensioned convolutional neural networks obtain three convolutional layers by the connection in pond of the convolution of convolution-pond-again-again, from first three Individual convolutional layer extracts the wild feature of different feeling, and these features are merged in a manner of cascading merging, then by two Density map corresponding to convolutional layer output；
The loss function L (θ) of the multiple dimensioned convolutional neural networks is calculated, expression formula is as follows：
<mrow> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mi>I</mi> <mrow> <mn>2</mn> <mi>N</mi> </mrow> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <mo>|</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>,</mo> <mi>&amp;theta;</mi> <mo>)</mo> </mrow> <mo>-</mo> <mi>M</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>|</mo> <msubsup> <mo>|</mo> <mn>2</mn> <mn>2</mn> </msubsup> </mrow>
Wherein, N be input convolutional neural networks amount of images, xiFor the i-th width input picture of convolutional neural networks, M (xi) table Show the accurate density map matrix of the i-th width input picture；
After one initial parameter is set for convolutional neural networks, the loss L of input picture is calculated according to the accurate density map of reality (θ), the parameter of whole network is then updated in Optimized Iterative each time, until penalty values converge to a less value.
CN201711014291.XA 2017-10-25 2017-10-25 Image people counting method based on multiple dimensioned convolutional neural networks Pending CN107862261A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711014291.XA CN107862261A (en) 2017-10-25 2017-10-25 Image people counting method based on multiple dimensioned convolutional neural networks

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711014291.XA CN107862261A (en) 2017-10-25 2017-10-25 Image people counting method based on multiple dimensioned convolutional neural networks

Publications (1)

Publication Number Publication Date
CN107862261A true CN107862261A (en) 2018-03-30

Family

ID=61697892

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711014291.XA Pending CN107862261A (en) 2017-10-25 2017-10-25 Image people counting method based on multiple dimensioned convolutional neural networks

Country Status (1)

Country Link
CN (1) CN107862261A (en)

Cited By (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108876774A (en) * 2018-06-07 2018-11-23 浙江大学 A kind of people counting method based on convolutional neural networks
CN109166100A (en) * 2018-07-24 2019-01-08 中南大学 Multi-task learning method for cell count based on convolutional neural networks
CN109271960A (en) * 2018-10-08 2019-01-25 燕山大学 A kind of demographic method based on convolutional neural networks
CN109389044A (en) * 2018-09-10 2019-02-26 中国人民解放军陆军工程大学 Multi-scene crowd density estimation method based on convolutional network and multi-task learning
CN109447990A (en) * 2018-10-22 2019-03-08 北京旷视科技有限公司 Image, semantic dividing method, device, electronic equipment and computer-readable medium
CN109492615A (en) * 2018-11-29 2019-03-19 中山大学 Crowd density estimation method based on CNN low layer semantic feature density map
CN109543695A (en) * 2018-10-26 2019-03-29 复旦大学 General density people counting method based on multiple dimensioned deep learning
CN109558862A (en) * 2018-06-15 2019-04-02 广州深域信息科技有限公司 The people counting method and system of attention refinement frame based on spatial perception
CN109598220A (en) * 2018-11-26 2019-04-09 山东大学 A kind of demographic method based on the polynary multiple dimensioned convolution of input
CN109614941A (en) * 2018-12-14 2019-04-12 中山大学 A kind of embedded crowd density estimation method based on convolutional neural networks model
CN109919214A (en) * 2019-02-27 2019-06-21 南京地平线机器人技术有限公司 A kind of training method and training device of neural network model
CN110163060A (en) * 2018-11-07 2019-08-23 腾讯科技（深圳）有限公司 The determination method and electronic equipment of crowd density in image
CN110163057A (en) * 2018-10-29 2019-08-23 腾讯科技（深圳）有限公司 Object detection method, device, equipment and computer-readable medium
CN110598669A (en) * 2019-09-20 2019-12-20 郑州大学 Method and system for detecting crowd density in complex scene
CN110674704A (en) * 2019-09-05 2020-01-10 同济大学 Crowd density estimation method and device based on multi-scale expansion convolutional network
WO2020042169A1 (en) * 2018-08-31 2020-03-05 Intel Corporation 3d object recognition using 3d convolutional neural network with depth based multi-scale filters
CN110956057A (en) * 2018-09-26 2020-04-03 杭州海康威视数字技术股份有限公司 Crowd situation analysis method and device and electronic equipment
CN111027554A (en) * 2019-12-27 2020-04-17 创新奇智(重庆)科技有限公司 System and method for accurately detecting and positioning commodity price tag characters
CN111144398A (en) * 2018-11-02 2020-05-12 银河水滴科技（北京）有限公司 Target detection method, target detection device, computer equipment and storage medium
CN111144329A (en) * 2019-12-29 2020-05-12 北京工业大学 Light-weight rapid crowd counting method based on multiple labels
CN111191667A (en) * 2018-11-15 2020-05-22 天津大学青岛海洋技术研究院 Crowd counting method for generating confrontation network based on multiple scales
CN111209892A (en) * 2020-01-19 2020-05-29 浙江中创天成科技有限公司 Crowd density and quantity estimation method based on convolutional neural network
CN111242036A (en) * 2020-01-14 2020-06-05 西安建筑科技大学 Crowd counting method based on encoding-decoding structure multi-scale convolutional neural network
CN111428809A (en) * 2020-04-02 2020-07-17 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) Crowd counting method based on spatial information fusion and convolutional neural network
CN111460912A (en) * 2020-03-12 2020-07-28 南京理工大学 Dense crowd counting algorithm based on cascade high-resolution convolutional neural network
CN111611878A (en) * 2020-04-30 2020-09-01 杭州电子科技大学 Method for crowd counting and future people flow prediction based on video image
CN111626134A (en) * 2020-04-28 2020-09-04 上海交通大学 Dense crowd counting method, system and terminal based on hidden density distribution
CN111815665A (en) * 2020-07-10 2020-10-23 电子科技大学 Single image crowd counting method based on depth information and scale perception information
CN111951260A (en) * 2020-08-21 2020-11-17 苏州大学 Partial feature fusion based convolutional neural network real-time target counting system and method
CN112183728A (en) * 2020-09-29 2021-01-05 上海松鼠课堂人工智能科技有限公司 Learning strategy generation method and system based on deep learning
CN112287873A (en) * 2020-11-12 2021-01-29 广东恒电信息科技股份有限公司 Judicial service early warning system
CN112767316A (en) * 2020-12-31 2021-05-07 山东师范大学 Crowd counting method and system based on multi-scale interactive network
CN113706529A (en) * 2021-10-28 2021-11-26 鲁东大学 Method, system and device for counting abalone in seedling stage by using convolutional neural network
CN113887536A (en) * 2021-12-06 2022-01-04 松立控股集团股份有限公司 Multi-stage efficient crowd density estimation method based on high-level semantic guidance

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104992223A (en) * 2015-06-12 2015-10-21 安徽大学 Intensive population estimation method based on deep learning
CN105528589A (en) * 2015-12-31 2016-04-27 上海科技大学 Single image crowd counting algorithm based on multi-column convolutional neural network
US20160259980A1 (en) * 2015-03-03 2016-09-08 Umm Al-Qura University Systems and methodologies for performing intelligent perception based real-time counting
CN106203331A (en) * 2016-07-08 2016-12-07 苏州平江历史街区保护整治有限责任公司 A kind of crowd density evaluation method based on convolutional neural networks
CN106326937A (en) * 2016-08-31 2017-01-11 郑州金惠计算机系统工程有限公司 Convolutional neural network based crowd density distribution estimation method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160259980A1 (en) * 2015-03-03 2016-09-08 Umm Al-Qura University Systems and methodologies for performing intelligent perception based real-time counting
CN104992223A (en) * 2015-06-12 2015-10-21 安徽大学 Intensive population estimation method based on deep learning
CN105528589A (en) * 2015-12-31 2016-04-27 上海科技大学 Single image crowd counting algorithm based on multi-column convolutional neural network
CN106203331A (en) * 2016-07-08 2016-12-07 苏州平江历史街区保护整治有限责任公司 A kind of crowd density evaluation method based on convolutional neural networks
CN106326937A (en) * 2016-08-31 2017-01-11 郑州金惠计算机系统工程有限公司 Convolutional neural network based crowd density distribution estimation method

Cited By (52)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108876774A (en) * 2018-06-07 2018-11-23 浙江大学 A kind of people counting method based on convolutional neural networks
CN109558862A (en) * 2018-06-15 2019-04-02 广州深域信息科技有限公司 The people counting method and system of attention refinement frame based on spatial perception
CN109558862B (en) * 2018-06-15 2023-04-07 拓元(广州)智慧科技有限公司 Crowd counting method and system based on attention thinning framework of space perception
CN109166100A (en) * 2018-07-24 2019-01-08 中南大学 Multi-task learning method for cell count based on convolutional neural networks
WO2020042169A1 (en) * 2018-08-31 2020-03-05 Intel Corporation 3d object recognition using 3d convolutional neural network with depth based multi-scale filters
US11880770B2 (en) 2018-08-31 2024-01-23 Intel Corporation 3D object recognition using 3D convolutional neural network with depth based multi-scale filters
CN109389044A (en) * 2018-09-10 2019-02-26 中国人民解放军陆军工程大学 Multi-scene crowd density estimation method based on convolutional network and multi-task learning
CN109389044B (en) * 2018-09-10 2021-11-23 中国人民解放军陆军工程大学 Multi-scene crowd density estimation method based on convolutional network and multi-task learning
CN110956057A (en) * 2018-09-26 2020-04-03 杭州海康威视数字技术股份有限公司 Crowd situation analysis method and device and electronic equipment
CN109271960B (en) * 2018-10-08 2020-09-04 燕山大学 People counting method based on convolutional neural network
CN109271960A (en) * 2018-10-08 2019-01-25 燕山大学 A kind of demographic method based on convolutional neural networks
CN109447990B (en) * 2018-10-22 2021-06-22 北京旷视科技有限公司 Image semantic segmentation method and device, electronic equipment and computer readable medium
CN109447990A (en) * 2018-10-22 2019-03-08 北京旷视科技有限公司 Image, semantic dividing method, device, electronic equipment and computer-readable medium
CN109543695B (en) * 2018-10-26 2023-01-06 复旦大学 Population-density population counting method based on multi-scale deep learning
CN109543695A (en) * 2018-10-26 2019-03-29 复旦大学 General density people counting method based on multiple dimensioned deep learning
CN110163057A (en) * 2018-10-29 2019-08-23 腾讯科技（深圳）有限公司 Object detection method, device, equipment and computer-readable medium
CN110163057B (en) * 2018-10-29 2023-06-09 腾讯科技（深圳）有限公司 Object detection method, device, equipment and computer readable medium
CN111144398A (en) * 2018-11-02 2020-05-12 银河水滴科技（北京）有限公司 Target detection method, target detection device, computer equipment and storage medium
CN110163060B (en) * 2018-11-07 2022-12-23 腾讯科技（深圳）有限公司 Method for determining crowd density in image and electronic equipment
CN110163060A (en) * 2018-11-07 2019-08-23 腾讯科技（深圳）有限公司 The determination method and electronic equipment of crowd density in image
CN111191667A (en) * 2018-11-15 2020-05-22 天津大学青岛海洋技术研究院 Crowd counting method for generating confrontation network based on multiple scales
CN111191667B (en) * 2018-11-15 2023-08-18 天津大学青岛海洋技术研究院 Crowd counting method based on multiscale generation countermeasure network
CN109598220A (en) * 2018-11-26 2019-04-09 山东大学 A kind of demographic method based on the polynary multiple dimensioned convolution of input
CN109492615A (en) * 2018-11-29 2019-03-19 中山大学 Crowd density estimation method based on CNN low layer semantic feature density map
CN109614941A (en) * 2018-12-14 2019-04-12 中山大学 A kind of embedded crowd density estimation method based on convolutional neural networks model
CN109614941B (en) * 2018-12-14 2023-02-03 中山大学 Embedded crowd density estimation method based on convolutional neural network model
CN109919214A (en) * 2019-02-27 2019-06-21 南京地平线机器人技术有限公司 A kind of training method and training device of neural network model
CN109919214B (en) * 2019-02-27 2023-07-21 南京地平线机器人技术有限公司 Training method and training device for neural network model
CN110674704A (en) * 2019-09-05 2020-01-10 同济大学 Crowd density estimation method and device based on multi-scale expansion convolutional network
CN110598669A (en) * 2019-09-20 2019-12-20 郑州大学 Method and system for detecting crowd density in complex scene
CN111027554B (en) * 2019-12-27 2023-05-23 创新奇智(重庆)科技有限公司 Commodity price tag text accurate detection positioning system and positioning method
CN111027554A (en) * 2019-12-27 2020-04-17 创新奇智(重庆)科技有限公司 System and method for accurately detecting and positioning commodity price tag characters
CN111144329A (en) * 2019-12-29 2020-05-12 北京工业大学 Light-weight rapid crowd counting method based on multiple labels
CN111242036B (en) * 2020-01-14 2023-05-09 西安建筑科技大学 Crowd counting method based on multi-scale convolutional neural network of encoding-decoding structure
CN111242036A (en) * 2020-01-14 2020-06-05 西安建筑科技大学 Crowd counting method based on encoding-decoding structure multi-scale convolutional neural network
CN111209892A (en) * 2020-01-19 2020-05-29 浙江中创天成科技有限公司 Crowd density and quantity estimation method based on convolutional neural network
CN111460912B (en) * 2020-03-12 2022-10-28 南京理工大学 Dense crowd counting algorithm based on cascade high-resolution convolution neural network
CN111460912A (en) * 2020-03-12 2020-07-28 南京理工大学 Dense crowd counting algorithm based on cascade high-resolution convolutional neural network
CN111428809A (en) * 2020-04-02 2020-07-17 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) Crowd counting method based on spatial information fusion and convolutional neural network
CN111626134B (en) * 2020-04-28 2023-04-21 上海交通大学 Dense crowd counting method, system and terminal based on hidden density distribution
CN111626134A (en) * 2020-04-28 2020-09-04 上海交通大学 Dense crowd counting method, system and terminal based on hidden density distribution
CN111611878B (en) * 2020-04-30 2022-07-22 杭州电子科技大学 Method for crowd counting and future people flow prediction based on video image
CN111611878A (en) * 2020-04-30 2020-09-01 杭州电子科技大学 Method for crowd counting and future people flow prediction based on video image
CN111815665B (en) * 2020-07-10 2023-02-17 电子科技大学 Single image crowd counting method based on depth information and scale perception information
CN111815665A (en) * 2020-07-10 2020-10-23 电子科技大学 Single image crowd counting method based on depth information and scale perception information
CN111951260A (en) * 2020-08-21 2020-11-17 苏州大学 Partial feature fusion based convolutional neural network real-time target counting system and method
CN112183728A (en) * 2020-09-29 2021-01-05 上海松鼠课堂人工智能科技有限公司 Learning strategy generation method and system based on deep learning
CN112287873A (en) * 2020-11-12 2021-01-29 广东恒电信息科技股份有限公司 Judicial service early warning system
CN112767316A (en) * 2020-12-31 2021-05-07 山东师范大学 Crowd counting method and system based on multi-scale interactive network
CN113706529B (en) * 2021-10-28 2022-01-28 鲁东大学 Method, system and device for counting abalone in seedling stage by using convolutional neural network
CN113706529A (en) * 2021-10-28 2021-11-26 鲁东大学 Method, system and device for counting abalone in seedling stage by using convolutional neural network
CN113887536A (en) * 2021-12-06 2022-01-04 松立控股集团股份有限公司 Multi-stage efficient crowd density estimation method based on high-level semantic guidance

Similar Documents

Publication Publication Date Title
CN107862261A (en) Image people counting method based on multiple dimensioned convolutional neural networks
CN110210551B (en) Visual target tracking method based on adaptive subject sensitivity
CN109977812B (en) Vehicle-mounted video target detection method based on deep learning
CN107368831B (en) English words and digit recognition method in a kind of natural scene image
CN111612807B (en) Small target image segmentation method based on scale and edge information
CN104217214B (en) RGB D personage&#39;s Activity recognition methods based on configurable convolutional neural networks
CN105701508B (en) Global local optimum model and conspicuousness detection algorithm based on multistage convolutional neural networks
CN106682696B (en) The more example detection networks and its training method refined based on online example classification device
CN103984959B (en) A kind of image classification method based on data and task-driven
CN103955702B (en) SAR image terrain classification method based on depth RBF network
CN107368845A (en) A kind of Faster R CNN object detection methods based on optimization candidate region
CN110059581A (en) People counting method based on depth information of scene
CN109447033A (en) Vehicle front obstacle detection method based on YOLO
CN107506761A (en) Brain image dividing method and system based on notable inquiry learning convolutional neural networks
CN111783782A (en) Remote sensing image semantic segmentation method fusing and improving UNet and SegNet
CN107481188A (en) A kind of image super-resolution reconstructing method
CN109886066A (en) Fast target detection method based on the fusion of multiple dimensioned and multilayer feature
CN103984948B (en) A kind of soft double-deck age estimation method based on facial image fusion feature
CN111882620B (en) Road drivable area segmentation method based on multi-scale information
CN111382686B (en) Lane line detection method based on semi-supervised generation confrontation network
CN105160310A (en) 3D (three-dimensional) convolutional neural network based human body behavior recognition method
CN109409240A (en) A kind of SegNet remote sensing images semantic segmentation method of combination random walk
CN105528589A (en) Single image crowd counting algorithm based on multi-column convolutional neural network
CN106815604A (en) Method for viewing points detecting based on fusion of multi-layer information
CN106897681A (en) A kind of remote sensing images comparative analysis method and system

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20180330