Download | - View accepted manuscript: Fractal geometry and multimedia retrieval: a theoretical framework (PDF, 648 KiB)
|
---|
Author | Search for: Paquet, Eric1; Search for: Viktor, Herna L. |
---|
Affiliation | - National Research Council of Canada. Information and Communication Technologies
|
---|
Format | Text, Article |
---|
Conference | 3D Image Processing (3DIP) and Applications 2012, San Francisco Airport, Burlingame, California, 22-26 January 2012 |
---|
Subject | Bayesian; distance; distribution; fractal; geometry; indexing; Lagrangian; multimedia; object; probability; retrieval |
---|
Abstract | This paper presents a theoretical analysis of the occurrence of fractal geometry within index spaces and discusses the impact for multimedia retrieval. Firstly, we explain how to detect the presence of such a fractal geometry. Then, with the fractal hypothesis in hand, we analyze the impact of this geometry when calculating the distance between indexes and searching for similar multimedia objects. We demonstrate that it is possible to define probability distributions in fractal index space and we present an approach to calculate them. Practical consequences are discussed, with particular emphasis to multimedia retrieval methods based on Bayesian analysis. |
---|
Publication date | 2012-02 |
---|
Publisher | SPIE |
---|
In | |
---|
Series | |
---|
Language | English |
---|
Peer reviewed | Yes |
---|
NPARC number | 20255956 |
---|
Export citation | Export as RIS |
---|
Report a correction | Report a correction (opens in a new tab) |
---|
Record identifier | 5ad6b425-48bd-4c37-94fd-597aa8bdf79d |
---|
Record created | 2012-07-10 |
---|
Record modified | 2020-04-21 |
---|