Your browser does not support JavaScript!

Home    Περιβάλλον Ανάκτησης Εικόνων Βασισμένο σε Βιολογικά Εμπνευσμένη Αρχιτεκτονική  

Results - Details

Add to Basket
[Add to Basket]
Identifier uch.csd.msc//2002dimitriadis
Title Περιβάλλον Ανάκτησης Εικόνων Βασισμένο σε Βιολογικά Εμπνευσμένη Αρχιτεκτονική
Alternative Title An Image Retrieval Platform Based on a Biologically Inspired Architecture
Author Δημητριάδης, Σωκράτης
Thesis advisor Ορφανουδάκης, Στέλιος, 1948-2005
Abstract The rapid evolution of computers and telecommunications leads to the continuous increase of the volume of data. It is also a fact that the digital image constitutes a form of data with a rich content and an abundance of information. Given that humans have a natural competence of image perception while the computational approach is still falling short to the respective biological, we realize the need for efficient and intelligent techniques of image retrieval. In this research we study the issue of content-based image retrieval and we develop a corresponding platform with a biologically inspired architecture. The prospective applications extend from Internet search engines and the media to robotics and the medical diagnosis. The CBIR system we developed uses a number of autonomous software agents, each one undertaking a different primitive visual feature independently and in parallel. Most of the agents deal with features such as color, texture, shape and various combinations among them. The images along with the representations of all their features are stored in a database for a later retrieval, exploration and experimentation. The partial results that each agent extracts are fed into a voting system that supports several voting methodologies in order to obtain an overall similarity decision. The system is platform independent and can be easily customized to the requirements of the experiments. Furthermore, it provides many capabilities of extension and embedment for both new agents and voting methods.
Language Greek
Issue date 2002-12-05
Date available 2002-11-26
Collection   School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
  Type of Work--Post-graduate theses
Views 593

Digital Documents
No preview available

Download document
View document
Views : 12