Robust multiple cameras pedestrian detection with multi-view Bayesian network

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 视频内容安全

	Robust multiple cameras pedestrian detection with multi-view Bayesian network
	Peng PX(彭佩玺)1 ; Tian YH(田永鸿)1; Wang YW(王耀威)1; Li J(李甲)2; Huang TJ(黄铁军)1
发表期刊	Pattern Recognition
	2015
卷号	48 期号:5 页码:1760-1772
摘要	Multi-camera pedestrian detection is the challenging problem in the field of surveillance video analysis. However, existing approaches may produce "phantoms" (i.e., fake pedestrians) due to the heavy occlusions in real surveillance scenario, while calibration errors and the diverse heights of pedestrians may also heavily decrease the detection performance. To address these problems, this paper proposes a robust multiple cameras pedestrian detection approach with multi-view Bayesian network model (MvBN). Given the preliminary results obtained by any multi-view pedestrian detection method, which are actually comprised of both real pedestrians and phantoms, the MvBN is used to model both the occlusion relationship and the homography correspondence between them in all camera views. As such, the removal of phantoms can be formulated as an MvBN inference problem. Moreover, to reduce the influence of the calibration errors and keep robust to the diverse heights of pedestrians, a height-adaptive projection (HAP) method is proposed to further improve the detection performance by utilizing a local search process in a small neighborhood of heights and locations of the detected pedestrians. Experimental results on four public benchmarks show that our method outperforms several state-of-the-art algorithms remarkably and demonstrates high robustness in different surveillance scenes. HighlightsA multi-view Bayesian network is proposed to model pedestrian candidates and their occlusion relationships in all views.A parameter learning algorithm is developed for MvBN by using a set of auxiliary, real-valued, and continuous variables.A height-adaptive projection is proposed to make the final detection robust to synthesis noises and calibration errors.Our approach is recognized as the best performer in five PETS evaluations from 2009 to 2013.
关键词	Pedestrian Detection Multiple Cameras Multi-view Model Bayesian Inference Height Adaptive Projection
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/20216
专题	多模态人工智能系统全国重点实验室_视频内容安全
作者单位	1.北京大学 2.北京航空航天大学
推荐引用方式 GB/T 7714	Peng PX,Tian YH,Wang YW,et al. Robust multiple cameras pedestrian detection with multi-view Bayesian network[J]. Pattern Recognition,2015,48(5):1760-1772.
APA	Peng PX,Tian YH,Wang YW,Li J,&Huang TJ.(2015).Robust multiple cameras pedestrian detection with multi-view Bayesian network.Pattern Recognition,48(5),1760-1772.
MLA	Peng PX,et al."Robust multiple cameras pedestrian detection with multi-view Bayesian network".Pattern Recognition 48.5(2015):1760-1772.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
1000014133162.pdf（16022KB）	期刊论文	作者接受稿	开放获取	CC BY-NC-SA	浏览下载