Υλοποίηση συστήματος ανάκτησης εικόνων με βάση το περιεχόμενο

Βόγιας, Κωνσταντίνος Ι.

Master Thesis

Author

Βόγιας, Κωνσταντίνος Ι.

Date

2012-06-25

Abstract

Digital Images are a rich source of valuable information. The extraction of such information may lead to crucial conclusions regarding the content of each image used. Image recognition is possible by using image metadata or by processing its digital content. The systems that recognize images based on their content are known as CBIR (Content-Based Image Retrieval) and are designed to identify images based on their content and not on the metadata associated with (Lew, Nicu, Djeraba, & Ramesh, 2006). One of the techniques used in these systems, combine traditional image analysis algorithms for low level features extraction and data mining techniques such as classification, clustering and association rules extraction. After this, an image pattern is exported which represents the results of image mining. In this work our goal is to develop a CBIR system which integrates a pattern management and comparison system (Pattern Miner) with an image database. Furthermore, we will try to add semantics to the system for the semantic image classification and retrieval using a relevant ontology.