Fast algorithms for computing the largest empty rectangle

  • Authors:
  • A. Aggarwal;S. Suri

  • Affiliations:
  • IBM T. J. Watson Center, P. O. Box 218, Yorktown Heights, New York;Department of Computer Science, The Johns Hopkins University, Baltimore, MD

  • Venue:
  • SCG '87 Proceedings of the third annual symposium on Computational geometry
  • Year:
  • 1987

Quantified Score

Hi-index 0.01

Visualization

Abstract

We provide two algorithms for solving the following problem: Given a rectangle containing n points, compute the largest-area and the largest-perimeter subrectangles with sides parallel to the given rectangle that lie within this rectangle and that do not contain any points in their interior. For finding the largest-area empty rectangle, the first algorithm takes &Ogr;(n log3 n) time and &Ogr;(n) memory space and it simplifies the algorithm given by Chazelle, Drysdale and Lee which takes &Ogr;(n log3 n) time but &Ogr;(n log n) storage. The second algorithm for computing the largest-area empty rectangle is more complicated but it only takes &Ogr;(n log2 n) time and &Ogr;(n) memory space. The two algorithms for computing the largest-area rectangle can be modified to compute the largest-perimeter rectangle in &Ogr;(n log2 n) and &Ogr;(n log n) time, respectively. Since &OHgr;(n log n) is a lower bound on time for computing the largest-perimeter empty rectangle, the second algorithm for computing such a rectangle is optimal within a multiplicative constant.