1.淮北师范大学 物理与电子信息学院,安徽 淮北 235000
单 巍,男,讲师,现从事图像处理、模式识别方面的研究。E-mail: shanweifox@foxmail.com
扫 描 看 全 文
单巍, 王江涛, 方振国, 等. 基于深度学习的可见光图像中行人检测方法[J]. 武汉大学学报(理学版), 2021,67(2):127-135.
SHAN Wei, WANG Jiangtao, FANG Zhenguo, et al. Pedestrian Detection Method Based on Deep Learning in Visible Light Image[J]. J Wuhan Univ (Nat Sci Ed), 2021,67(2):127-135.
单巍, 王江涛, 方振国, 等. 基于深度学习的可见光图像中行人检测方法[J]. 武汉大学学报(理学版), 2021,67(2):127-135. DOI: 10.14188/j.1671-8836.2020.0115.
SHAN Wei, WANG Jiangtao, FANG Zhenguo, et al. Pedestrian Detection Method Based on Deep Learning in Visible Light Image[J]. J Wuhan Univ (Nat Sci Ed), 2021,67(2):127-135. DOI: 10.14188/j.1671-8836.2020.0115.
相对传统的行人检测技术,基于深度学习的行人检测技术具有压倒性的优势,然而由于深度卷积网络规模庞大,需要专用的处理器,限制了行人检测系统的推广。针对上述问题,提出一种网络规模适中的深度卷积网络模型,在保证检测精度的前提下提高检测模型的普适性。以低维度的浅层卷积神经网络为基础,分别从网络层数、感受野大小和特征图3个角度出发搜索最优的网络结构,并通过有指导的实验评估确定最终的网络参数。在Daimler行人数据库上进行实验,结果表明,本文建立的网络不但网络规模适中,而且具备良好的检测性能。在Daimler、MIT、INRIA等行人数据库上进行的交叉实验验证了依本文方法建立网络的鲁棒性,表明其具有推广能力。
Compared with traditional pedestrian detection technology, deep learning-based pedestrian detection technology has overwhelming advantages. However, due to the large scale of deep convolution network, the need for dedicated processors limits the promotion of pedestrian detection system. To solve this problem, this paper presents a deep convolution network model with moderate network size to improve the universality of the detection model while ensuring the detection accuracy. Based on the low-dimensional shallow convolution neural network, the optimal network structure is searched from three perspectives: the number of layers of the network, the size of the sensing field and the feature map, and the final network parameters are determined through guided experimental evaluation. On Daimler pedestrian database, the results show that the network structure designed in this paper not only has a moderate network scale, but also has a good detection performance. On Daimler, MIT and INRIA pedestrian databases, the cross-over experiments show that our network is robust and can be extended.
深度学习卷积神经网络可见光图像检测率
deep learningconvolutional neural networkvisible light imagedetection rate
陈震,杨小平,张聪炫,等.基于R-MI-rényi测度的可见光与红外图像配准[J].电子测量与仪器学报,2018,32(1):1-8. DOI: 10.13382/j.jemi.2018.01.001http://dx.doi.org/10.13382/j.jemi.2018.01.001.
CHEN Z, YANG X P, ZHANG C X,et al. Infrared and visible image registration based on R-MI-rényi measurement [J]. Journal of Electronic Measurement and Instrumentation, 2018,32(1):1-8. DOI: 10.13382/j.jemi.2018.01.001(Chhttp://dx.doi.org/10.13382/j.jemi.2018.01.001(Ch).
叶华,朱明旱,王日兴.红外和可见光图像互补融合的运动目标检测方法[J].红外技术,2015,37(8):648-654.DOI:10.11846/j.issn.1001_8891.201508004http://dx.doi.org/10.11846/j.issn.1001_8891.201508004.
YE H, ZHU M H, WANG R X. Moving target detection method based on complementary fusion of infrared and visible image [J]. Infrared Technology, 2015,37(8):648-654. DOI:10.11846/j.issn.1001_8891.201508004(Chhttp://dx.doi.org/10.11846/j.issn.1001_8891.201508004(Ch).
KIM S, JI Y, LEE K B. An effective sign language learning with object detection based ROI segmentation [C]// IEEE International Conference on Robotic Computing. Piscataway:IEEE, 2018:17190142. DOI: 10.1109/IRC.2018.00069http://dx.doi.org/10.1109/IRC.2018.00069.
SUN W, ZHAO H, JIN Z. A visual attention based ROI detection method for facial expression recognition [J]. Neuro Computing, 2018,296:12-22.
高文,汤洋,朱明.复杂背景下目标检测的级联分类器算法研究[J]. 物理学报, 2014, 63(9):94204. DOI:10.7498/aps.63.094204http://dx.doi.org/10.7498/aps.63.094204.
GAO W,TANG Y,ZHU M. Research on cascade classifier for target detection in complex background [J]. Acta Physica Sinica, 2014, 63(9):094204. DOI:10.7498/aps.63.094204(Chhttp://dx.doi.org/10.7498/aps.63.094204(Ch).
唐春晖.一种基于梯度方向直方图的俯视行人的检测方法[J].模式识别与人工智能,2015,28(1):19-26. DOI:10.16451/j.cnki.issn1003-6059.201501003http://dx.doi.org/10.16451/j.cnki.issn1003-6059.201501003.
TANG C H. A detection method of looking down pedestrian based on gradient direction histogram [J]. Pattern Recognition and Artificial Intelligence, 2015,28(1): 19-26. DOI:10.16451/j.cnki.issn1003-6059.201501003(Chhttp://dx.doi.org/10.16451/j.cnki.issn1003-6059.201501003(Ch).
徐渊,许晓亮,李才年,等.结合SVM分类器与HOG特征提取的行人检测[J].计算机工程,2016,42(1):56-60+65. DOI:10.3969/j.issn.1000-3428.2016.01.011http://dx.doi.org/10.3969/j.issn.1000-3428.2016.01.011.
XU Y,XU X L,LI C N,et al. Pedestrian detection combining SVM classifier and hog feature extraction [J]. Computer Engineering, 2016,42(1): 56-60 + 65. DOI:10.3969/j.issn.1000-3428.2016.01.011(Chhttp://dx.doi.org/10.3969/j.issn.1000-3428.2016.01.011(Ch).
邹冲,蔡敦波,刘莹,等.组合SVM分类器在行人检测中的研究[J].计算机科学,2017,44(S1):188-191. DOI:CNKI:SUN:JSJA.0.2017-S1-043http://dx.doi.org/CNKI:SUN:JSJA.0.2017-S1-043.
ZOU C,CAI D B,LIU Y,et al. Research on combined SVM classifier in pedestrian detection [J]. Computer Science, 2017,44(S1): 188-191. DOI:CNKI:SUN:JSJA.0.2017-S1-043(Chhttp://dx.doi.org/CNKI:SUN:JSJA.0.2017-S1-043(Ch).
曲永宇,刘清,郭建明,等.基于HOG和颜色特征的行人检测[J].武汉理工大学学报,2011,33(4):134-138.DOI:CNKI:SUN:WHGY.0.2011-04-032http://dx.doi.org/CNKI:SUN:WHGY.0.2011-04-032.
QU Y Y,LIU Q,GUO J M,et al. Pedestrian detection based on HOG and color features [J]. Journal of Wuhan University of Technology, 2011,33(4): 134-138. DOI:CNKI:SUN:WHGY.0.2011-04-032(Chhttp://dx.doi.org/CNKI:SUN:WHGY.0.2011-04-032(Ch).
樊春年,杜卫平,刘艳荣.基于HOG特征结合AdaBoost算法的行人检测[J].自动化技术与应用,2018,37(7):89-91. 10.3969/j.issn.1003-7241.2018.07.022http://dx.doi.org/10.3969/j.issn.1003-7241.2018.07.022
FAN C N,DU W P,LIIU Y R. Pedestrian detection based on hog features and AdaBoost algorithm [J]. Techniques of Automation and Applications, 2018,37(7): 89-91(Ch). 10.3969/j.issn.1003-7241.2018.07.022http://dx.doi.org/10.3969/j.issn.1003-7241.2018.07.022
HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks [J]. Science, 2006,313(5786):504-507. DOI: 10.1126/science.1127647http://dx.doi.org/10.1126/science.1127647.
ZENG X Y, OUYANG W L, WANG X G. Multi-stage contextual deep learning for pedestrian detection[C]// Proceedings of International Conference on Computer Vision. Piscataway:IEEE,2013: 14132232. DOI: 10.1109/ICCV.2013.22http://dx.doi.org/10.1109/ICCV.2013.22.
TIAN Y Y, LUO P , WANG X G , et al. Pedestrian detection aided by deep learning semantic tasks [C]// IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE,2015:5079-5087. DOI: 10.1109/CVPR.2015.7299143http://dx.doi.org/10.1109/CVPR.2015.7299143.
张汇,杜煜,宁淑荣,等.基于Faster RCNN的行人检测方法[J].传感器与微系统,2019,38(2):147-149+153.DOI:10.13873/J.1000-9787(2019)02-0147-03http://dx.doi.org/10.13873/J.1000-9787(2019)02-0147-03.
ZHANG H,DU Y,NING S R,et al. Pedestrian detection method based on Faster RCNN [J]. Transducer and Microsystem Technologies, 2019,38(2): 147-149 + 153. DOI:10.13873/J.1000-9787(2019)02-0147-03(Chhttp://dx.doi.org/10.13873/J.1000-9787(2019)02-0147-03(Ch).
郝旭政,柴争义.一种改进的深度残差网络行人检测方法[J].计算机应用研究,2019,36(5):1569-1572+1584.
HAO X Z,CHAI Z Y. An improved pedestrian detection method based on depth residual network [J]. Application Research of Computers, 2019,36(5): 1569-1572 + 1584(Ch).
LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11):2278-2324. DOI: 10.1109/5.726791http://dx.doi.org/10.1109/5.726791.
崔鹏,范志旭.基于域鉴别网络和域自适应的行人重识别[J].光电子·激光,2019,30(6):632-639.
CUI P, FAN Z X. Pedestrian re-identification based on domain discriminative network and domain adaptation [J].Journal of Optoelectronics·Laser, 2019,30(6):632-639(Ch).
费建超,芮挺,周遊,等.基于梯度的多输入卷积神经网络[J].光电工程,2015,42(3):33-38. DOI:10.3969/j.issn.1003-501X.2015.03.006http://dx.doi.org/10.3969/j.issn.1003-501X.2015.03.006.
FEI J C,RUI T,ZHOU Y,et al. Multi input convolution neural network based on gradient [J]. Opto⁃Electronic Engineering, 2015,42(3): 33-38. DOI:10.3969/j.issn.1003-501X.2015.03.006(Chhttp://dx.doi.org/10.3969/j.issn.1003-501X.2015.03.006(Ch).
DALAL N, TRIGGS B. Histograms of oriented gradients for human detection [C]// IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2005: 886-893. DOI: 10.1109/CVPR.2005.177http://dx.doi.org/10.1109/CVPR.2005.177.
0
浏览量
45
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构