|本期目录/Table of Contents|

[1]庞彦伟,尚楚博,何宇清.基于尺度不变特征和位置先验的行人检测算法[J].天津大学学报(自然科学版),2017,(09):946-952.[doi:10.11784/tdxbz201608024]
 Pang Yanwei,Shang Chubo,He Yuqing.Pedestrian Detection Algorithm Based on Scale Invariant Features and Prior Position Information[J].Journal of Tianjin University,2017,(09):946-952.[doi:10.11784/tdxbz201608024]
点击复制

基于尺度不变特征和位置先验的行人检测算法()
分享到:

《天津大学学报(自然科学版)》[ISSN:0493-2137/CN:12-1127/N]

卷:
期数:
2017年09
页码:
946-952
栏目:
电气自动化与信息工程
出版日期:
2017-09-22

文章信息/Info

Title:
Pedestrian Detection Algorithm Based on Scale Invariant Features and Prior Position Information
文章编号:
0493-2137(2017)09-0946-07
作者:
庞彦伟 尚楚博 何宇清
天津大学电气自动化与信息工程学院,天津 300072
Author(s):
Pang Yanwei Shang Chubo He Yuqing
School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China
关键词:
行人检测 尺度不变 位置先验 回归模型
Keywords:
pedestrian detection scale invariant prior position information regression model
分类号:
TP391.4
DOI:
10.11784/tdxbz201608024
文献标志码:
A
摘要:
基于相邻和非相邻特征(NNNF)行人检测算法, 提出了一种方法来解决行人特征对尺度变化敏感的问题以及窗口误检的问题.首先, 在NNNF基础上, 设计了一种尺度不变的特征.由韦伯定理启发, 该特征表示为两个相邻或非相邻区域的差分值与这两个区域特征和的比值.这种新的特征具有很强的尺度不变性.此外, 还提出了基于行人位置先验的上下文信息, 作为一种简单有效的后处理方法.在行人场景中, 行人的高度与位置存在一定的映射关系.利用SVM(support vector machine)训练了行人高度关于行人位置的回归模型.该模型能有效地滤除那些行人高度与位置信息不符合回归模型的检测窗口.实验表明, 相比于NNNF-L2和NNNF-L4, 本文提出的方法在Caltech数据库的检测性能分别提高了2.90% 和2.28% .同时, 本文提出的方法也在所有基于非深度网络的行人检测方法中具有最好的检测性能, 平均漏检率为14.56% .
Abstract:
Based on neighbouring and non-neighbouring features(NNNF),a method was proposed to address challenges in pedestrian detection,such as features being sensitive to scale changes and the existence of a large number of false positives. First,based on NNNF,a scale invariant feature was proposed. This feature is computed as the ratio of difference to sum between two neighbouring or non-neighbouring regions’ local feature value,inspired by the Weber law in experimental psychology. The new feature is strong scale invariant. Second,a contextual information based on the prior position information was proposed,and this method serves as simple and effective post-processing approach. In the scene of pedestrians,there is some function relationship between the height and position of pedestrian. The support vector machine(SVM)was utilized to train a regression model of height in relation to position. And this regression model can effectively filter out those false positives whose height and position are inconsistent with the regression model. Furthermore,experimental results on Caltech pedestrian dataset show that the proposed method gives an average 2.90% and 2.28% improvement over the NNNF-L2 and NNNF-L4,respectively. Most importantly,among all the methods without using CNN,the proposed method achieves state-of-the-art performance(i. e.,14.56% average miss rate on Caltech dataset).

参考文献/References:

[1] Benenson R, Omran M, Hosang J, et al. Ten years of pedestrian detection, what have we learned?[C]// Proceedings of European Conference on Computer Vision. Zurich, Switzerland, 2014:613-627.
[2] Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA, 2014:580-587.
[3] Hosang J H, Omran M, Benenson R, et al. Taking a deeper look at pedestrians[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA, 2015:1-9.
[4] Sermanet P, Kavukcuoglu K, Chintala S, et al. Pedestrian detection with unsupervised multi-stage feature learning[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Portland, USA, 2013:3626-3633.
[5] Dollar P, Tu Zhouwen, Peron P, et al. Integral channel features[C]// Proceedings of British Machine Vision Conference. London, UK, 2009:1-11.
[6] Appel R, Fuchs T, Dollar P, et al. Quickly boosting decision trees-oruning underachieving features early[C]// Proceedings of International Conference on Machine Learning. Atlanta, USA, 2013:594-602.
[7] Dalal N, Triggs B. Histograms of oriented gradients for human detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. San Diego, USA, 2005:886-893.
[8] Dollar P, Appel R, Belongie S, et al. Fastest feature pyramids for object detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(8):1532-1545.
[9] Nam W, Dollar P, Han J H. Local decorrelation for improved pedestrian detection[C]// Proceedings of Advances in Neural Information Processing Systems. Montreal, Canada, 2014:424-432.
[10] Zhang Shanshan, Bauckhage C, Cremers A B. Informed Haar-like features improve pedestrian detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA, 2014:947-954.
[11] Zhang Shanshan, Benenson R, Schiele B. Filtered channel features for pedestrian detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA, 2015:1751-1760.
[12] Cao Jiale, Pang Yanwei, Li Xuelong. Pedestrian detection inspired by appearance constancy and shape symmetry[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA, 2016:5538-5551.
[13] Hoiem D, Efros A, Hebert M. Putting objects in perspective[J]. International Journal of Computer Vision, 2008, 80(1):3-15.
[14] Park D, Ramanan D, Fowlkes C. Multiresolution models for object detection[C]// Proceedings of European Conference on Computer Vision. Zurich, Switzerland, 2010:241-254.
[15] Weber E H, Hering E. Tastsinn und Gemeingefuhl[M]. Leipzig:Hering, 1905.
[16] Dollar P, Wojek C, Schiele B, et al. Pedestrian detection:An evaluation of the state of the art[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(4):743-761.
[17] Luo Ping, Tian Yonglong, Wang Xiaogang, et al. Switchable deep network for pedestrian detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA, 2014:899-906.
[18] Paisitkriangkrai S, Shen Chunhua, van den Hengel A. Strengthening the effectiveness of pedestrian detection with spatially pooled features[C]// Proceedings of European Conference on Computer Vision. Zurich, Switzerland, 2014:546-561.
[19] Tian Yonglong, Luo Ping, Wang Xiaogang, et al. Pe-
destrian detection aided by deep learning semantic tasks[C] //Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA, 2015:5079-5087.
[20] Tian Yonglong, Luo Ping, Wang Xiaogang, et al. Deep learning strong parts for pedestrian detection[C]// Proceedings of IEEE International Conference on Computer Vision. Santiago, Chile, 2016:1904-1912.
[21] Cai Zhaowei, Saberian M, Vasconcelos N. Learning complexity-aware cascades for deep pedestrian detection [C] //Proceedings of IEEE International Conference on Computer Vision. Zurich, Switzerland, 2016:1-10.
[22] Zhang Shanshan, Benenson R, Omran M, et al. How far are we from solving pedestrian detection?[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA, 2016:1259-1267.

相似文献/References:

[1]金志刚,赵明昕,张瑞,等.基于双目视觉的聚合积分通道行人检测优化算法[J].天津大学学报(自然科学版),2016,(12):1225.[doi:10.11784/tdxbz201506016]
 Jin Zhigang,Zhao Mingxin,Zhang Rui,et al.Optimized Aggregated Channel Features Pedestrian Detection Algorithm Based on Binocular Vision[J].Journal of Tianjin University,2016,(09):1225.[doi:10.11784/tdxbz201506016]

备注/Memo

备注/Memo:
收稿日期: 2016-08-16; 修回日期: 2016-11-22.
作者简介: 庞彦伟(1976—), 男, 博士, 教授.
通讯作者: 庞彦伟, pyw@tju.edu.cn.
基金项目: 国家自然科学基金资助项目(61472274).
Supported by the National Natural Science Foundation of China(No. 61472274).
更新日期/Last Update: 2017-09-10