Image retrieval using fuzzy segmentation and a graph matching technique

We describe the design and implementation of a system for content-based image retrieval (CBIR). The system uses structural description by labelled graphs. When a graph representation is used, the comparison task is performed by a graph matching technique. Like all CBIR systems, the one proposed here has two components: a segmentation component which uses an efficient fuzzy clustering algorithm and a matching component which employs a new Inexact graph matching algorithm.