robust maximum likelihood estimation by sparse bundle adjustment using the l1 norm

ISCAS OpenIR

	robust maximum likelihood estimation by sparse bundle adjustment using the l1 norm
	Dai Zhijun; Zhang Fengjun; Wang Hongan
	2012
会议名称	2012 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2012
会议录名称	Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
页码	1672-1679
会议日期	June 16, 2012 - June 21, 2012
会议地点	Providence, RI, United states
收录类别	EI ; ISTP
ISSN	1063-6919
ISBN	9781467312264
部门归属	(1) Intelligence Engineering Lab Institute of Software Chinese Academy of Sciences China; (2) State Key Lab of Computer Science Institute of Software Chinese Academy of Sciences China
摘要	Sparse bundle adjustment is widely used in many computer vision applications. In this paper, we propose a method for performing bundle adjustments using the L1 norm. After linearizing the mapping function in bundle adjustment on its first order, the kernel step is to compute the L1 norm equations. Considering the sparsity of the Jacobian matrix in linearizing, we find two practical methods to solve the L1 norm equations. The first one is an interior-point method, which transfer the original problem to a problem of solving a sequence of L2 norm equations, and the second one is a decomposition method which uses the differentiability of linear programs and represents the optimal updating of parameters of 3D points by the updating variables of camera parameters. The experiments show that the method performs better for both synthetically generated and real data sets in the presence of outliers or Laplacian noise compared with the L2 norm bundle adjustment, and the method is efficient among the state of the art L1 minimization methods. © 2012 IEEE.; Sparse bundle adjustment is widely used in many computer vision applications. In this paper, we propose a method for performing bundle adjustments using the L1 norm. After linearizing the mapping function in bundle adjustment on its first order, the kernel step is to compute the L1 norm equations. Considering the sparsity of the Jacobian matrix in linearizing, we find two practical methods to solve the L1 norm equations. The first one is an interior-point method, which transfer the original problem to a problem of solving a sequence of L2 norm equations, and the second one is a decomposition method which uses the differentiability of linear programs and represents the optimal updating of parameters of 3D points by the updating variables of camera parameters. The experiments show that the method performs better for both synthetically generated and real data sets in the presence of outliers or Laplacian noise compared with the L2 norm bundle adjustment, and the method is efficient among the state of the art L1 minimization methods. © 2012 IEEE.
关键词	Jacobian Matrices Maximum Likelihood Estimation
主办者	IEEE
学科领域	Computer Science ; Engineering
语种	英语
内容类型	会议论文
URI标识	http://ir.iscas.ac.cn/handle/311060/15789
专题	中国科学院软件研究所
推荐引用方式 GB/T 7714	Dai Zhijun,Zhang Fengjun,Wang Hongan. robust maximum likelihood estimation by sparse bundle adjustment using the l1 norm[C],2012:1672-1679.