28.02.2013 Views

Bio-medical Ontologies Maintenance and Change Management

Bio-medical Ontologies Maintenance and Change Management

Bio-medical Ontologies Maintenance and Change Management

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

134 L. Stanescu, D. Dan Burdescu, <strong>and</strong> M. Brezovan<br />

6 Conclusions<br />

The chapter presents a series of aspects in the field of multimedia <strong>medical</strong> databases.<br />

These databases are the result of structuring the alphanumeric <strong>and</strong> imagistic data<br />

gathered in large quantities in the patient investigation <strong>and</strong> diagnosis processes.<br />

We have also presented the necessity of creating <strong>and</strong> managing the multimedia<br />

<strong>medical</strong> databases <strong>and</strong> the advantages of using these operations in order to increase<br />

<strong>medical</strong> act efficiency.<br />

A big part of <strong>medical</strong> data is stored in DICOM files that can’t be seen directly<br />

on a computer. As a result, we have given attention to this problem by presenting<br />

ours original algorithms for extracting the alphanumeric <strong>and</strong> imagistic data from<br />

this type of files in order to be integrated in the multimedia <strong>medical</strong> databases.<br />

Also, the problem of DICOM viewers has been presented. These tools have the<br />

ability to display <strong>and</strong> apply certain operations on the information from DICOM<br />

files, being very useful in the <strong>medical</strong> activity.<br />

Taking into consideration that a database is created for querying <strong>and</strong> obtaining<br />

with accuracy <strong>and</strong> speed the information requested by the user, the chapter treats<br />

in great detail the problem of content-based visual query. This type of query is<br />

applied on multimedia databases <strong>and</strong> can be combined with text-based simple<br />

query.<br />

As a rule, the content-based visual query takes into consideration the primitive<br />

characteristics automatically extracted from images: color, texture or shape.<br />

We have presented in detail the notion of color <strong>and</strong> algorithms for automated<br />

extraction of the gray-level or color information from <strong>medical</strong> images <strong>and</strong> also<br />

aspects of color texture characteristics extracting, algorithms <strong>and</strong> experiments on<br />

color images from digestive area.<br />

Another important aspect in <strong>medical</strong> applications is the automated segmentation<br />

of images. For segmenting the color images from digestive area, we have implemented<br />

the color set back-projection algorithm. The chapter presents in detail<br />

ours original implementations of this method <strong>and</strong> the results of the content-based<br />

region query on color, size <strong>and</strong> spatial extent on a database with color <strong>medical</strong><br />

images.<br />

The presented experiments have been realized with the help of MIR (Medical<br />

Image Retrieval) – a software tool for creating, managing <strong>and</strong> content-based visual<br />

querying of the multimedia <strong>medical</strong> databases with color images from the digestive<br />

area.<br />

These experiments were realized along with <strong>medical</strong> experts from the two biggest<br />

university hospitals from Craiova <strong>and</strong> with financial support from Romanian<br />

Academy <strong>and</strong> National University Research Council.<br />

In the content-based visual query of the <strong>medical</strong> databases <strong>and</strong> automated segmentation<br />

of <strong>medical</strong> images areas, there are wide researches, a big variety of<br />

techniques applied on various imaging modalities, taking into consideration many<br />

systems of the human body <strong>and</strong> pathologies.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!