Tomographie ultrasonore dédiée à la détection du cancer du sein

Abstract : We studied two types of ultrasound tomography corresponding to two acquisition configurations : qualitative reflection tomography and quantitative transmission tomography. Both are based on the hypothesis of weakly heterogeneous medium in order to use the Born approximation in reflection tomography and the straight ray approximation in velocity and absorption transmission tomography.
Qualitative tomography : we developped a filtered elliptical backprojection algorithm based on an elliptical Radon transform and to a near-field extension of the classical Fourier projection-slice theorem. This inversion method was studied experimentally on phantoms of agar-agar gel and paraffin immersed in water. Moreover, in order to be closer from the practitioner's condition, we developed 2-D anatomic breast ductal computer phantoms for ultrasonic imaging (tomography and echography). These phantoms allowed us to simulate and compare both imaging methods.
Quantitative tomography for tissue characterisation : For the sound speed parameter, we compared a layer stripping method developed by Ferrière et al (2003) and a classical fan beam tomography, both of which are based on a straight ray assumption. Several numerical tests and one experimental test showed the limits of the layer stripping method to reproduce accurately the objects. Concerning the absorption parameter, we showed numerically that it is essential to take into account the scattering effects, that are generally neglected in the case of soft tissues like those of breast. We studied the error introduced by the scattering phenomenon in the absorption estimation frequency domain method. We proposed a correction method for scattering abberations which gives good results.
Complete list of metadatas

Cited literature [174 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00139036
Contributor : Emilie Franceschini <>
Submitted on : Wednesday, March 28, 2007 - 10:15:54 PM
Last modification on : Monday, March 4, 2019 - 2:04:03 PM
Long-term archiving on : Tuesday, April 6, 2010 - 11:12:34 PM

Identifiers

  • HAL Id : tel-00139036, version 1

Citation

Emilie Franceschini. Tomographie ultrasonore dédiée à la détection du cancer du sein. Acoustique [physics.class-ph]. Université de Provence - Aix-Marseille I, 2006. Français. ⟨tel-00139036⟩

Share

Metrics

Record views

591

Files downloads

760