Dynamic Weighted Voronoi Diagrams and Weighted Minimax Matching of Two Corresponding Point.pdfVIP

Dynamic Weighted Voronoi Diagrams and Weighted Minimax Matching of Two Corresponding Point Sets Keiko Imai Dept. of Information and System Engineering, Chuo University Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan E-mail: imai@ise.chuo-u.ac.jp Hiroshi Imai Dept. of Information Science, University of Tokyo Hongo, Bunkyo-ku, Tokyo 113-0033, Japan Dedicated to Professor Masao Iri on the occasion of his 65th birthday Abstract A weighted geometric fitting problem between two corresponding sets of points is to minimize the maximum weighted distance between two corresponding pairs of points by translating and rotating one set to the other set. For this weighted geometric fitting problem, dynamic weighted Voronoi diagrams have been applied to obtain an almost cubic algorithm. In this paper, we show a new reduction of the problem to the two-dimensional Davenport-Schinzel sequences, and provide a much simpler proof for the almost cubic bound. The technique used for this purpose can be applied to more general cases. Keywords: Dynamic Voronoi diagram, Davenport-Schinzel sequence, envelopes, computa- tional geometry 1 Introduction The minimax geometric fitting problem between two similar sets of points is considered in [18]. The problem arises in an industrial robot attaching a pin-grid-array type LSI to a board by using visual sensors. The weighted problem can be briefly stated as follows. Given two sets S = {s = (x , y ) | j = 1, · · · , n} and T = {t = (u , v ) | j = 1, · · · , n}