Knowledge base: Warsaw University of Technology

Settings and your account

Back

Multiscale methods of data representation in medical imaging

Rafał Jóźwiak

Abstract

Presented thesis concerns the problem of multiscale representation methods of medical images. Its power emerges in terms of observed progress in the field of medical informatics. Overcoming existing limitations requires an adjustment of all forms of medical image processing to unique features of diagnostic content. It becomes necessary to understand both the medical and technical conditions of diagnostic content variability as well as specific limitations of perception and interpretation. It becomes possible to model the composition of image content in order to establish the key diagnostic components and their essential characteristics. The aim of authors’ thesis was to propose efficient algorithmic solutions using multiscale representation of the source data. Key features of ultiscale representations such as hierarchyal, locality, compactness and selectivity of content description allow to adapt effective form of the representation to objective features of diagnostic content. In the thesis created or co-created by author, practical algorithmic solutions for radiography image compression and computer-aided diagnosis in selected diagnostic tasks are presented. The obtained results show the usefulness of proposed methods, conforming experimantally proved thesis.
Record ID
WUTae18639f30fe40bbb6baa18201750e17
Diploma type
Doctor of Philosophy
Author
Title in Polish
Wieloskalowe metody reprezentacji tresci diagnostycznej w obrazach medycznych
Title in English
Multiscale methods of data representation in medical imaging
Language
(pl) Polish
Certifying Unit
Faculty of Electronics and Information Technology (FEIT)
Discipline
electronics / (technology domain) / (technological sciences)
Status
Finished
Defense Date
16-12-2014
Supervisor
Pages
262
Keywords in English
imaging data, medical informatics
Abstract in English
Presented thesis concerns the problem of multiscale representation methods of medical images. Its power emerges in terms of observed progress in the field of medical informatics. Overcoming existing limitations requires an adjustment of all forms of medical image processing to unique features of diagnostic content. It becomes necessary to understand both the medical and technical conditions of diagnostic content variability as well as specific limitations of perception and interpretation. It becomes possible to model the composition of image content in order to establish the key diagnostic components and their essential characteristics. The aim of authors’ thesis was to propose efficient algorithmic solutions using multiscale representation of the source data. Key features of ultiscale representations such as hierarchyal, locality, compactness and selectivity of content description allow to adapt effective form of the representation to objective features of diagnostic content. In the thesis created or co-created by author, practical algorithmic solutions for radiography image compression and computer-aided diagnosis in selected diagnostic tasks are presented. The obtained results show the usefulness of proposed methods, conforming experimantally proved thesis.
Thesis file
Request a WCAG compliant version
Reviews
Request a WCAG compliant version

Uniform Resource Identifier
https://repo.pw.edu.pl/info/phd/WUTae18639f30fe40bbb6baa18201750e17/
URN
urn:pw-repo:WUTae18639f30fe40bbb6baa18201750e17

Confirmation
Are you sure?
Report incorrect data on this page