Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature
Crowd counting is an important research topic in the field of computer vision. The multi-column convolution neural network (MCNN) has been used in this field and achieved competitive performance. However, when the crowd distribution is uneven, the accuracy of crowd counting based on the MCNN still n...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8755826/ |
| _version_ | 1818323418752221184 |
|---|---|
| author | Zhi Liu Yue Chen Bo Chen Linan Zhu Du Wu Guojiang Shen |
| author_facet | Zhi Liu Yue Chen Bo Chen Linan Zhu Du Wu Guojiang Shen |
| author_sort | Zhi Liu |
| collection | DOAJ |
| description | Crowd counting is an important research topic in the field of computer vision. The multi-column convolution neural network (MCNN) has been used in this field and achieved competitive performance. However, when the crowd distribution is uneven, the accuracy of crowd counting based on the MCNN still needs to be improved. In order to adapt to uneven crowd distributions, crowd global density feature is taken into account in this paper. The global density features are extracted and added to the MCNN through the cascaded learning method. Because some detailed features during the down-sampling process will be lost in the MCNN and it will affect the accuracy of the density map, an improved MCNN structure is proposed. In this paper, the max pooling is replaced by max-ave pooling to keep more detailed features and the deconvolutional layers are added to restore the lost details in the down-sampling process. The experimental results in the UCF_CC_50 dataset and the ShanghaiTech dataset show that the proposed method has higher accuracy and stability. |
| first_indexed | 2024-12-13T11:12:23Z |
| format | Article |
| id | doaj.art-a4b88060db5741f1af1229b5c880b5fc |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-13T11:12:23Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-a4b88060db5741f1af1229b5c880b5fc2022-12-21T23:48:43ZengIEEEIEEE Access2169-35362019-01-017887898879810.1109/ACCESS.2019.29268818755826Crowd Counting Method Based on Convolutional Neural Network With Global Density FeatureZhi Liu0https://orcid.org/0000-0001-8320-820XYue Chen1https://orcid.org/0000-0002-1446-1460Bo Chen2https://orcid.org/0000-0001-8889-3744Linan Zhu3https://orcid.org/0000-0002-7451-4421Du Wu4https://orcid.org/0000-0002-4002-0837Guojiang Shen5https://orcid.org/0000-0003-1064-1250College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, ChinaCollege of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, ChinaCollege of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, ChinaCollege of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, ChinaCollege of Computer Science and Technology, Jilin University, Changchun, ChinaCollege of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, ChinaCrowd counting is an important research topic in the field of computer vision. The multi-column convolution neural network (MCNN) has been used in this field and achieved competitive performance. However, when the crowd distribution is uneven, the accuracy of crowd counting based on the MCNN still needs to be improved. In order to adapt to uneven crowd distributions, crowd global density feature is taken into account in this paper. The global density features are extracted and added to the MCNN through the cascaded learning method. Because some detailed features during the down-sampling process will be lost in the MCNN and it will affect the accuracy of the density map, an improved MCNN structure is proposed. In this paper, the max pooling is replaced by max-ave pooling to keep more detailed features and the deconvolutional layers are added to restore the lost details in the down-sampling process. The experimental results in the UCF_CC_50 dataset and the ShanghaiTech dataset show that the proposed method has higher accuracy and stability.https://ieeexplore.ieee.org/document/8755826/Global density featuredeep learningconvolutional neural networkcrowd counting |
| spellingShingle | Zhi Liu Yue Chen Bo Chen Linan Zhu Du Wu Guojiang Shen Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature IEEE Access Global density feature deep learning convolutional neural network crowd counting |
| title | Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature |
| title_full | Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature |
| title_fullStr | Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature |
| title_full_unstemmed | Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature |
| title_short | Crowd Counting Method Based on Convolutional Neural Network With Global Density Feature |
| title_sort | crowd counting method based on convolutional neural network with global density feature |
| topic | Global density feature deep learning convolutional neural network crowd counting |
| url | https://ieeexplore.ieee.org/document/8755826/ |
| work_keys_str_mv | AT zhiliu crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature AT yuechen crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature AT bochen crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature AT linanzhu crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature AT duwu crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature AT guojiangshen crowdcountingmethodbasedonconvolutionalneuralnetworkwithglobaldensityfeature |