The concept of content-based visual image retrieval system in the experimental medical database
-
1
University of West Bohemia, Czechia
The Content-based visual information retrieval (CBVIR) has been one of the most growing research area over the last few years. The reason is steadily increasing amount of multimedia, especially visual data in wide range of nowadays professional activities. The CBVIR technics offer intelligent similarity data search, which can be used if the comparison among many cases are needed. In medical applications the recent diagnostic and therapy procedures usually involve work with the latest technical equipments and imaging devices. The new graphic diagnostic methods include for instance perfusion computed tomography CTP, CT angiography or diffusion weighted magnetic resonance MR DWI. This approach produces huge amounts of medical images for each patient and study case, generally in the international standard Digital Imaging and Communications in Medicine (DICOM). The CBVIR doorway should be helpful in the future work with this great amount of stored medical images for the clinical staff as well as the researchers and scientists.
The common CBVIR system engine architecture contains some basic functional modules €“ storage and access methods, visual feature extraction, distance measures and similarity calculations and user interface and interaction methods. Each incoming case (image or array of images) is analysed, the distinguished visual features are extracted and are compared with the features of stored images. The visual features are classified into primitive features (color, grey level, shape or texture), logical features like an object identity and abstract features such as significance of scenes pictured. In the medical applications, the color and grey level features are often neglected, especially due to the lack of the contrast reference point in the radiology images. More frequently the texture and shape features are applied. The technics for texture identification use for instance Canny operators, invariant moments, scale-space filtering, Gabor filters, wavelets and Markov texture characteristics or Fourier descriptors for shape characterization. Also the segmentation of the incoming image into smaller parts is investigated. In our contemporary research we try to develop the open web-based multimedia database system of complex neurological information for each real medical case. This system will first store the medical images in DICOM format and more information about the current study. One of the objectives is to design and set up CBVIR engine for our experimental database. The beginning basic proposal of the CBVIR in context of our database system is presented. The retrieval engine will use comparison of histogram vectors and the optimal method for similarity calculations must be selected. The basic visual features of the one case´s images will be combined with the other information from the database, such as from the textual section or form section. Then the aggregate histogram vector will be created, stored in the database and can be used in the further similarity study retrieval. The presented concept of the CBVIR system will be used in future work, mainly in design process of our experimental database system.
Conference:
Neuroinformatics 2009, Pilsen, Czechia, 6 Sep - 8 Sep, 2009.
Presentation Type:
Poster Presentation
Topic:
Neuroimaging
Citation:
Polivka
J and
Kleckova
J
(2019). The concept of content-based visual image retrieval system in the experimental medical database.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2009.
doi: 10.3389/conf.neuro.11.2009.08.023
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
21 May 2009;
Published Online:
09 May 2019.
*
Correspondence:
Jiri Polivka, University of West Bohemia, Pilsen, Czechia, polivkaj@kiv.zcu.cz