Region-based CBIR in GIS with local space filling curves to spatial representation

Region-based CBIR in GIS with local space filling curves to spatial representation

^aInstitut d’Electronique Fondamentale, Université de Paris-Sud 11, Bât 220, 91405 Orsay Cedex, France

^bDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai 400 076, India

^cDepartment of Electrical and Computer Engineering, Air Force Institute of Technology, Wright-Patterson AFB, OH 45433-7765, USA

Available online 30 September 2005.

Abstract

In this paper we present a region-based retrieval method for satellite images using motif co-occurrence matrix (MCM) in conjunction with spatial relationships. Each image is decomposed into coherent segments, MCM is computed for each region and the spatial relationship among them are evaluated by using a *-tree representation. The image is represented by an attributed relational graph (ARG) where nodes contain the visual feature (MCM) and edges represent spatial relationship. Principal component analysis show the usefulness of MCM as a feature.

Keywords: Motif co-occurrence matrix; Peano scan; Region-based image retrieval; Spatial relationship; *-tree; Fuzzy c-means clustering

Article Outline

1. Introduction

2. Feature extraction

2.1. Motif co-occurrence matrix

2.2. Region-based spatial relationship

2.2.1. Definitions

2.2.2. Distance and angle in a tree

3. MCM properties

3.1. Illumination invariance

3.2. Invariance to resolution changes

3.3. Separability of classes

4. Segmentation

5. Image retrieval

6. Experiments and results

7. Conclusion

References