基于深度卷积神经网络的人群密度估计方法

被引:9
作者
谭智勇 [1 ]
袁家政 [1 ,2 ]
刘宏哲 [1 ]
李青 [1 ]
机构
[1] 北京市信息服务工程重点实验室
[2] 北京成像技术高精尖创新中心
基金
北京市自然科学基金;
关键词
人群密度估计; 图像分块; 深度卷积神经网络;
D O I
暂无
中图分类号
TP183 [人工神经网络与计算]; TP391.41 [];
学科分类号
080203 ;
摘要
人群密度自动估计作为人群控制和管理的方法,是当前视频监控中的一个重要研究领域。现有的方法通过提取复杂的特征来进行人群密度估计,由于人群遮挡、透视效果和环境复杂等条件限制,难以满足实际应用中的需求,而深度卷积神经网络在特征学习上具有较强的能力。提出了一种基于深度卷积神经网络DCNN(Deep Convolution Neural Network)的方法来进行自然场景下人群密度估计。首先,为了消除摄像机透视效果,以图像中行人身高作为尺度基准,将图像分成多个子图像块。其次,设计一种新的深度卷积神经网络结构,利用多种不同的卷积核提取人群图像的深层次特征进行人群密度估计。实验结果证明该方法在自然场景下人群密度估计具有良好的稳定性和鲁棒性。
引用
收藏
页码:130 / 136
页数:7
相关论文
共 18 条
  • [1] Estimation of Crowd Density Based on Adaptive LBP[J] . Yue Li,Tao Zou,Peng Chen. &nbspAdvanced Materials Research . 2014 (998)
  • [2] Estimation of Crowd Density Based on Wavelet and Support Vector Machine. Li Xiaohua,Shen Lansun,Li Huanqin. Transactions of the Institute of Measurement and Control . 2006
  • [3] CDES: A pixel-based crowd density estimation system for Masjid al-Haram[J] . Norhaida Hussain,Halimatul Saadiah Md. Yatim,Nor Liza Hussain,Jasy Liew Suet Yan,Fazilah Haron. &nbspSafety Science . 2011 (6)
  • [4] Deep ID3:face recognition with very deep neural networks. SUN Y,LIANG D,WANG X G,et al. Computer Science . 2015
  • [5] Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. Girshick,Ross,Donahue,Jeff,Darrell,Trevor,et al. Computer Science . 2014
  • [6] Convolutional neural networks for speech recognition. Abdel-Hamid O,Mohamed A R,Jiang H,et al. ACM Transactions on Audio,Speech,and Language Processing . 2014
  • [7] Over Feat:Integrated Recognition,Localization and Detection using Convolutional Networks. Sermanet P,Eigen D,Zhang X,et al. . 2013
  • [8] Cross-scene crowd counting via deep convolutional neural networks. Zhang C,Li H,Wang X,et al. IEEE Conference on Computer Vision&Pattern Recognition . 2015
  • [9] Crowd monitoring using image processing. Davies, A.C.,Jia Hong Yin,Velastin, S.A. Electronics and Communication Engineering Journal . 1995
  • [10] Estimation of Crowd Density in Public Areas Based on Neural Network. Gyujin Kim,Taeki An,Moonhyun Kim. KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS . 2012