| | |
Summary: SIIM 2008 Abstract Submission Form
Descriptive Abstract
Email to: abstracts@siimweb.org
Abstract Title:
A Unifying Framework for Combining Content-Based Image Retrieval with Relational Database
Queries for Biomedical Applications
Topic:
TRIPTM
Preferred Presentation:
Paper Presentation (10-12 Mins)
Background:
Biomedical research scientists often conduct large-scale experiments involving multi-modality
data such as neuronal recordings, CT scans and fMRI images. The non-textual nature and
increasingly large amount of these data necessitates that they be retrieved not only via standard
relational database queries over the meta data (such as patient information, reason for
examination and date of exam), but also via measures of data similarity (such as image texture or
3D shape). Traditional relational database retrieval methods are designed for efficiency of
retrieval and are extremely effective for exact retrieval on alphanumeric data. Content-based
image retrieval (CBIR) methods were designed for retrieving images that are similar to a query
image according to one or more domain-dependent distance measures [1]. CBIR systems cannot
|
