Statistical Inference for Spatial Processes

This book introduces statistical inference in spatial statistics and its applications.

B. D. Ripley (Author)

9780521424202, Cambridge University Press

Paperback / softback, published 18 July 1991

160 pages, 50 b/w illus.
22.9 x 15.2 x 0.9 cm, 0.24 kg

'All in all, a beautiful introduction to this important area and highly recommended.' Mededelingen van het Wiskundig Genootschap

The study of spatial processes and their applications is an important topic in statistics and finds wide application particularly in computer vision and image processing. This book is devoted to statistical inference in spatial statistics and is intended for specialists needing an introduction to the subject and to its applications. One of the themes of the book is the demonstration of how these techniques give new insights into classical procedures (including new examples in likelihood theory) and newer statistical paradigms such as Monte-Carlo inference and pseudo-likelihood. Professor Ripley also stresses the importance of edge effects and of lack of a unique asymptotic setting in spatial problems. Throughout, the author discusses the foundational issues posed and the difficulties, both computational and philosophical, which arise. The final chapters consider image restoration and segmentation methods and the averaging and summarising of images. Thus, the book will find wide appeal to researchers in computer vision, image processing, and those applying microscopy in biology, geology and materials science, as well as to statisticians interested in the foundations of their discipline.

Introduction
1. Likelihood analysis for spatial Gaussian processes
2. Edge correction for spatial point processes
3. Parameter estimation for Gibbsian point processes
4. Modelling spatial images
5. Summarizing binary images.

Subject Areas: Probability & statistics [PBT]