New Complexity Bounds for Image Matching under Rotation and Scaling

  • Authors:
  • Christian Hundt;Maciej Liśkiewicz

  • Affiliations:
  • Institut für Informatik, Universität Rostock, Germany;Institut für Theoretische Informatik, Universität zu Lübeck, Germany

  • Venue:
  • CPM '09 Proceedings of the 20th Annual Symposium on Combinatorial Pattern Matching
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The problem of image matching under rotation is to find for two given digital images A and B a rotation that converts image A as close as possible to B . The research in combinatorial pattern matching led to a series of improving algorithms which commonly attack this problem by searching the complete set of all rotations of A . We present the first optimal algorithm of this kind, i.e, one that solves image matching in time for images of size n ×n . Subsequently, for image matching under compositions of rotation and scaling we show a new lower bound *** (n 6 / logn ) on the cardinality of , the set of rotated and scaled transformations of A . This bound almost matches the upper bound O (n 6).