Content-Based Image Retrieval with Cellular Automata - Cellular Automata in Image Processing and Geometry - page 141

Chapter 8
Content-Based Image Retrieval with Cellular
Automata

Lynette van Zijl

Abstract. In this contribution, the practical application of cellular automata (CA)
to content-based image retrieval is considered. A brief background on the standard
content-based image retrieval processes is given, followed by the basic CA mod-
els to achieve many of the processes throughout most of the content-based image
retrieval pipeline. The chapter concludes with a practical case study.

8.1 Introduction

Given an image, how can one retrieve a similar-looking image from a database of
images? The field of content-based image retrieval (CBIR) addresses exactly this
problem. This chapter considers the role that cellular automata (CA) can play in
implementing the many processes involved in the CBIR pipeline, particularly in the
feature analysis of the original image. Although the use of CA is well known in
image processing, little work has been done to integrate CA into the whole pro-
cessing pipeline for CBIR. Exceptions are the work by Van Zijl et al. [2, 32] and
Konstantinidis et al. [18].

8.2 Content-Based Image Retrieval: A Background

CBIR is a broad field, with many issues and perspectives at play. For a detailed
overview, the reader is referred to some of the well-known books [28] and sur-
veys [8] on the field. In this chapter, the focus is primarily on the use of cellular au-
tomata to achieve the implementations of many of the processes involved in CBIR.
Prior knowledge of cellular automata is assumed, as for example in [13].
In CBIR, the contents of a source image is used to find the set of nearest matching
images in a database of images. For example, given a picture of a bird, one would

Lynette van Zijl
Stellenbosch University, Stellenbosch, South Africa
e-mail: [e-mail]