Lq-Closest-Point to Affine Subspaces using the Generalized Weiszfeld Algorithm
نویسندگان
چکیده
This paper presents a method for finding an Lqclosest-point to a set of affine subspaces, that is a point for which the sum of the q-th power of orthogonal distances to all the subspaces is minimized, where 1 ≤ q < 2. We give a theoretical proof for the convergence of the proposed algorithm to a unique Lq minimum. The proposed method is motivated by the Lq Weiszfeld algorithm, an extremely simple and rapid averaging algorithm, that finds the Lq mean of a set of given points in a Euclidean space. The proposed algorithm is applied to the triangulation problem in computer vision by finding the Lq-closest-point to a set of lines in 3D. Our experimental results for the triangulation problem confirm that the Lq-closest-point method, for 1 ≤ q < 2, is more robust to outliers than the L2-closest-point method.
منابع مشابه
Pseudoframe multiresolution structure on abelian locally compact groups
Let $G$ be a locally compact abelian group. The concept of a generalized multiresolution structure (GMS) in $L^2(G)$ is discussed which is a generalization of GMS in $L^2(mathbb{R})$. Basically a GMS in $L^2(G)$ consists of an increasing sequence of closed subspaces of $L^2(G)$ and a pseudoframe of translation type at each level. Also, the construction of affine frames for $L^2(G)$ bas...
متن کاملGeneralized penetration depth computation based on kinematical geometry
The generalized penetration depth PD of two overlapping bodies X and Y is the distance between the given colliding position of X and the closest collision-free Euclidean copy Xε to X according to a distance metric. We present geometric optimization algorithms for the computation of PD with respect to an object-oriented metric S which takes the mass distribution of the moving body X into conside...
متن کاملBig Data Representation as High Dimensional Configuration
Consider a high-dimensional data set, such that for every data-point there is incomplete information. Each object in the data set represents a real entity, which models as a point in high-dimensional space. We assume that all real entities are embedded in the same space, which means they have the same dimension. We model the lack of information for a given object as affine subspace in Rd with d...
متن کاملA Generalized Krylov Subspace Method for ℓp-ℓq Minimization
This paper presents a new efficient approach for the solution of the lp-lq minimization problem based on the application of successive orthogonal projections onto generalized Krylov subspaces of increasing dimension. The subspaces are generated according to the iteratively reweighted least-squares strategy for the approximation of lp/lq-norms by weighted l2-norms. Computed image restoration exa...
متن کاملPoint-Based Statistical Shape Models with Probabilistic Correspondences and Affine EM-ICP
A fundamental problem when computing statistical shape models (SSMs) is the determination of correspondences between the instances. Often, homologies between points that represent the surfaces are assumed which might lead to imprecise mean shape and variation results. We present a novel algorithm based on the affine Expectation Maximization Iterative Closest Point (EM-ICP) registration method. ...
متن کامل