ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-09252019-145807


Tipo di tesi
Tesi di laurea magistrale
Autore
BARBIERI, DAVIDE
URN
etd-09252019-145807
Titolo
Data harmonization in mammography: the tissue equivalent representation
Dipartimento
FISICA
Corso di studi
FISICA
Relatori
relatore Prof.ssa Fantacci, Maria Evelina
relatore Prof. Oliva, Piernicola
commissario Prof. Mannella, Riccardo
commissario Prof. Fidecaro, Francesco
commissario Prof. Forti, Francesco
commissario Prof. Guadagnini, Enore
commissario Prof. Leporini, Dino
commissario Prof. Roddaro, Stefano
commissario Prof. Shore, Steven Neil
Parole chiave
  • machine learning applications
  • databases
  • data harmonization
  • artificial intelligence applications
  • mammography
  • medical physics
  • representation
Data inizio appello
16/10/2019
Consultabilità
Non consultabile
Data di rilascio
16/10/2089
Riassunto
In this master thesis I modelled the image formation in mammography to evaluate the thickness of a material of interest in a multi-material object.
I started by working on the characterization of the mammographic unit of both the x-ray tube and its detector.
The equalization process of the image has been approached with the characterization of the several parameters like the dark current, the pixel intensity response to exposure at different kVp. This process showed a significant improvement especially at higher level of exposures. The output of the tube has then been measured and used in order to calibrate the simulated spectra. This has made the characterization of the energy response of the system possible. The scattering component of the radiation has been empirically characterized with the η factor.
Finally, the computation of the EGT has then been approached with two different methods the first of which considered the pixel intensity while the second made energy considerations. Where early results showed a big discrepancy between the reconstructed and real thickness of the material the scattering correction was implemented.
File