Abstract Views :382 |
PDF Views:167
Authors
Affiliations
1 School of Computer Science and Electronic Engineering, University of Essex, Colchester, GB
2 Department of Computer Engineering, Sharif University of Technology, Tehran, IR
Source
AIRCC's International Journal of Computer Science and Information Technology, Vol 2, No 1 (2010), Pagination: 77-83
Abstract
This work presents a semi-automatic tool for content retrieval. In contrast to traditional content based image retrieval systems that work with entire images, the tool we have developed handles individual objects. The availability of a pool of pre-segmented objects found using region analysis allows the human behaviour of presegmentation to be replicated. To generate defined objects in the object pool, segmentation is performed using multidimensional Binary Partition Trees (BPTs). The tree structure uses colour, spatial frequency edge histograms to form semantically meaningful tree nodes. The BPTs can be intuitively browsed and are stored within XML documents for ease of access and analysis. To find an object a node from a query image is matched against the nodes of the BPT of the database image. These are matched according to a collection of MPEG-7 descriptors. Performance evaluation shows high quality segmentations and reliable retrieval results.
Keywords
Binary Partition Tree, Segmentation, Content Retrieval, Object Extraction, MPEG-7.
Full Text
