The University of Southampton
University of Southampton Institutional Repository

Content based image retrieval using scale space object trees

Content based image retrieval using scale space object trees
Content based image retrieval using scale space object trees

Semantic extraction from images is of growing interest in the field of content based retrieval. Classical algorithms are mainly limited to the use of features such as colour or texture estimated globally over the whole image, or the use of shape description for clearly extractable object boundaries. There is little provision for the searching of images based on any semantic notion, other than by the use of matching keywords which are manually associated with the images. We believe that it is the combination of a number of image properties, estimated locally for individual regions rather than globally, that will create the ability to recognise objects, and from there we can work towards bridging the semantic gap.

Using recent advances in scale-space decomposition techniques, it is possible to transform images into trees in scale space, which represent the topology of regions within the image. Using a sieve mechanism we can decompose an image into a discrete scale-space based on the pixel area of regions within the image. Extrema within the image are merged to create new extrema. The merge operations are represented by branches in a tree. The tree then represents the topology of extrema in an image.

We propose that complex objects can be found in the trees using a combination of subgraph isomorphism testing and conventional feature matching and hence provide a vehicle for achieving content-based retrieval and navigation for complex objects in images.  Explanations are given for the subgraph isomorphism techniques, and how they can be supplemented with feature matching to produce scores for feature-topology matches of images.  Finally, we will explain how the scale-trees could be used in combination with a semantic layer to achieve semantic recognition of scenes, by recognition of parts of the topology of a scale-tree as particular objects.

University of Southampton
Dupplaw, David Paul
4c4cc6a9-d4dd-461b-8f2d-4dcfc2f14345
Dupplaw, David Paul
4c4cc6a9-d4dd-461b-8f2d-4dcfc2f14345

Dupplaw, David Paul (2002) Content based image retrieval using scale space object trees. University of Southampton, Doctoral Thesis.

Record type: Thesis (Doctoral)

Abstract

Semantic extraction from images is of growing interest in the field of content based retrieval. Classical algorithms are mainly limited to the use of features such as colour or texture estimated globally over the whole image, or the use of shape description for clearly extractable object boundaries. There is little provision for the searching of images based on any semantic notion, other than by the use of matching keywords which are manually associated with the images. We believe that it is the combination of a number of image properties, estimated locally for individual regions rather than globally, that will create the ability to recognise objects, and from there we can work towards bridging the semantic gap.

Using recent advances in scale-space decomposition techniques, it is possible to transform images into trees in scale space, which represent the topology of regions within the image. Using a sieve mechanism we can decompose an image into a discrete scale-space based on the pixel area of regions within the image. Extrema within the image are merged to create new extrema. The merge operations are represented by branches in a tree. The tree then represents the topology of extrema in an image.

We propose that complex objects can be found in the trees using a combination of subgraph isomorphism testing and conventional feature matching and hence provide a vehicle for achieving content-based retrieval and navigation for complex objects in images.  Explanations are given for the subgraph isomorphism techniques, and how they can be supplemented with feature matching to produce scores for feature-topology matches of images.  Finally, we will explain how the scale-trees could be used in combination with a semantic layer to achieve semantic recognition of scenes, by recognition of parts of the topology of a scale-tree as particular objects.

Text
894743.pdf - Version of Record
Available under License University of Southampton Thesis Licence.
Download (25MB)

More information

Published date: 2002

Identifiers

Local EPrints ID: 464933
URI: http://eprints.soton.ac.uk/id/eprint/464933
PURE UUID: 4be3201c-1a0f-4d64-998c-1324c92fd3f2

Catalogue record

Date deposited: 05 Jul 2022 00:12
Last modified: 16 Mar 2024 19:50

Export record

Contributors

Author: David Paul Dupplaw

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×