Metoda analizy sekwencji chmur punktów na potrzeby lokalizacji i śledzenia wybranych fragmentów powierzchni kończyn dolnych człowieka

Marcin Rafał Witkowski

Abstract

The dissertation describes a new method for processing and analysis of 4D data for localizing and tracing of characteristic surface areas of human body to support physicians in patient’s health assessment. The most important step of the developed method is the analysis of distribution maps of parameters describing locally the surface shape. To fulfill the requirements of medical analysis two new shape parameters C1 and C2 were developed. Their advantages are short calculation time and high resistance to noise in the input data. Two methods - manual and automated - were developed for localization of characteristic areas. The automated method is based on the analysis of a two-dimensional histogram showing the occurrence frequency of specific C1 and C2 parameters values. In co-operation with medical doctors a feature vector was developed that describes medically important parameters of the characteristic areas. The areas are traced along the sequence time with application of a method using time prediction algorithm. All developed methods were implemented in the NogiProject application. It was used for the preliminary validation of the method with use of computer generated data as well as results of real measurements. The application features an easy-to-use user interface adjusted to medical staff requirements and allowing for three-dimensional visualization and user validation of the analysis process on every its stage. Thanks to this the NogiProject application was implemented at the Department of Sport Medicine of Tübingen Universitat (Germany) and the Department of Biomechanics at the Katholieke Universitat Leuven (Belgium). The internal architecture of the application allows for its further development. This allowed for development of two additional modules in response to the demands of medical community. One module is used in the assessment of thoracic wall deformation in pectus excavatum patients while the other allows for calculation of analytic Cobb angles in kyphosis/lordosis assessment. The modules are used at the Medical University of Warsaw for research and clinical examination.
Diploma typeDoctor of Philosophy
Author Marcin Rafał Witkowski (FM / IMPh)
Marcin Rafał Witkowski,,
- The Institute of Micromechanics and Photonics
Title in PolishMetoda analizy sekwencji chmur punktów na potrzeby lokalizacji i śledzenia wybranych fragmentów powierzchni kończyn dolnych człowieka
Languagepl polski
Certifying UnitFaculty of Mechatronics (FM)
Disciplinebiocybernetics and biomedical engineering / (technology domain) / (technological sciences)
Start date16-06-2006
Defense Date29-09-2010
End date13-10-2010
Supervisor Małgorzata Kujawińska (FM / IMPh)
Małgorzata Kujawińska,,
- The Institute of Micromechanics and Photonics

Internal reviewers Barbara Putz (FM / IACR)
Barbara Putz,,
- The Institute of Automatic Control and Robotics
External reviewers Andrzej Górecki
Andrzej Górecki,,
-

Juliusz Lech Kulikowski - [Polish Academy of Sciences (PAN)]
Juliusz Lech Kulikowski,,
-
- Polska Akademia Nauk
Pages156
Keywords in Englishengineering biomedical
Abstract in EnglishThe dissertation describes a new method for processing and analysis of 4D data for localizing and tracing of characteristic surface areas of human body to support physicians in patient’s health assessment. The most important step of the developed method is the analysis of distribution maps of parameters describing locally the surface shape. To fulfill the requirements of medical analysis two new shape parameters C1 and C2 were developed. Their advantages are short calculation time and high resistance to noise in the input data. Two methods - manual and automated - were developed for localization of characteristic areas. The automated method is based on the analysis of a two-dimensional histogram showing the occurrence frequency of specific C1 and C2 parameters values. In co-operation with medical doctors a feature vector was developed that describes medically important parameters of the characteristic areas. The areas are traced along the sequence time with application of a method using time prediction algorithm. All developed methods were implemented in the NogiProject application. It was used for the preliminary validation of the method with use of computer generated data as well as results of real measurements. The application features an easy-to-use user interface adjusted to medical staff requirements and allowing for three-dimensional visualization and user validation of the analysis process on every its stage. Thanks to this the NogiProject application was implemented at the Department of Sport Medicine of Tübingen Universitat (Germany) and the Department of Biomechanics at the Katholieke Universitat Leuven (Belgium). The internal architecture of the application allows for its further development. This allowed for development of two additional modules in response to the demands of medical community. One module is used in the assessment of thoracic wall deformation in pectus excavatum patients while the other allows for calculation of analytic Cobb angles in kyphosis/lordosis assessment. The modules are used at the Medical University of Warsaw for research and clinical examination.
Thesis file
Microsoft Word - MW - Doktorat - ALL18.pdf 13.57 MB
Citation count*5 (2020-09-18)

Get link to the record

Back
Confirmation
Are you sure?