Options
Binary Histograms for Resource Selection in Peer-to-Peer Media Retrieval
Blank, Daniel; Henrich, Andreas (2010): Binary Histograms for Resource Selection in Peer-to-Peer Media Retrieval, in: Martin Atzmueller, Benz. Dominik, Andreas Hotho, u. a. (Hrsg.), LWA 2010 : Lernen, Wissen & Adaptivität ; Workshop Proceedings, Kassel, S. 183–190.
Faculty/Chair:
Author:
Title of the compilation:
LWA 2010 : Lernen, Wissen & Adaptivität ; Workshop Proceedings
Editors:
Atzmueller, Martin
Benz. Dominik
Hotho, Andreas
Stumme, Gerd
Conference:
LWA 2010 : Lernen, Wissen & Adaptivität, October 4 - 6, 2010 ; Kassel
Publisher Information:
Year of publication:
2010
Pages:
Series ; Volume:
Kasseler Informatikschriften ; 2010,5
Language:
English
Abstract:
With the ever increasing amount of media data and collections on the world wide web and on private devices arises a strong need for adequate indexing and search techniques. Trends such as personal media archives, social networks, mobile devices with huge storage space and networks with high bandwidth capacities make distributed solutions and peer-to-peer (P2P) systems attractive. Here, resource selection can be applied to determine a ranking of promising resources based on descriptions of their content. Resources are contacted in ranked order to retrieve appropriate media items w.r.t. a user’s information need.
In this paper we apply and adapt resource descriptions in the form of binary histograms and corresponding selection techniques which were designed for low-dimensional spatial data to high-dimensional data in the context of contentbased image retrieval (CBIR). W.r.t. related work in distributed information retrieval, which is also discussed in this paper, a main characteristic of our approach are more space efficient resource descriptions. This makes them applicable for a wider range of application fields apart from the P2P domain.
In this paper we apply and adapt resource descriptions in the form of binary histograms and corresponding selection techniques which were designed for low-dimensional spatial data to high-dimensional data in the context of contentbased image retrieval (CBIR). W.r.t. related work in distributed information retrieval, which is also discussed in this paper, a main characteristic of our approach are more space efficient resource descriptions. This makes them applicable for a wider range of application fields apart from the P2P domain.
GND Keywords: ; ;
Histogramm
Ressourcenmanagement
Information Retrieval
Keywords:
Binary Histograms
DDC Classification:
RVK Classification:
Type:
Conferenceobject
Activation date:
April 30, 2014
Versioning
Question on publication
Permalink
https://fis.uni-bamberg.de/handle/uniba/4737