Detecting features in spatial point processes with clutter via local indicators of spatial association

J. Mateu, G. Lorenzo, E. Porcu

Research output: Contribution to journalArticlepeer-review

13 Scopus citations

Abstract

We consider the problem of detecting features of general shape in spatial point processes in the presence of substantial clutter. Our goal is to remove clutter from images where one or several features are present and have to be detected. We use a method based on local indicators of spatial association (LISA) functions, particularly on the development of a local version of the product density which is a second-order characteristic of spatial point processes. The classification method is built upon a stochastic version of the EM algorithm (SEM). This method can be applied without user input about the number or shapes of the regions. Our proposal, compared with the kth nearest-neighbor technique, is tested through simulated examples yielding high detection and low false-positive rates. Two real case studies of connective loose tissues in human organs and earthquakes are also presented.

Original languageBritish English
Pages (from-to)968-990
Number of pages23
JournalJournal of Computational and Graphical Statistics
Volume16
Issue number4
DOIs
StatePublished - Dec 2007

Keywords

  • EM and SEM algorithms
  • Feature detection
  • LISA functions
  • Nearest-neighbor distances
  • Product density
  • Spatial point processes

Fingerprint

Dive into the research topics of 'Detecting features in spatial point processes with clutter via local indicators of spatial association'. Together they form a unique fingerprint.

Cite this