Skip to Main content Skip to Navigation

Matrix approach for ultrasound imaging and quantification

Abstract : Ultrasound imaging relies on two major assumptions. First, the medium is considered as homogeneous with a constant speed of sound. Second, the back-scattered wave-field only contains singly-scattered echoes. Nonetheless, the speed of sound varies greatly in different tissues. These fluctuations give rise to a distortion of the incident and back-scattered wave-fronts. Moreover, multiple scattering events can also occur between the scatterers of the medium. This multiple scattering contribution manifests itself as an incoherent background noise in the RF signal. Those two undesirable effects, namely aberrations and multiple scattering, thus lead to a loss of resolution and contrast in the ultrasound image.Conventional ultrasound imaging techniques rely on arrays of transducers that can be individually controlled to emit or receive ultrasonic waves. State-of-the-art ultrasound images are based on a confocal method that consists in a double focusing, both in transmit and in receive, on each point of the medium corresponding to one pixel of the image. In this thesis, we propose a matrix approach of ultrasound imaging that basically consists in splitting the locations of the transmit and receive focal spots. This process gives access to the impulse responses between virtual transducers located within the medium at each pixel location. This set of responses form a so-called focused reflection matrix that contains all the available information on the medium under investigation. Besides describing all the current ultrasound imaging methods under a matrix formalism, matrix imaging is able to take up several challenges: (i) quantify and enhance the ultrasound image quality via a local focusing criterion and a matrix aberration correction; (ii) develop novel quantitative imaging modes by building maps of the speed-of-sound and of a multiple-scattering-rate that may constitute relevant biomarkers for ultrasound diagnosis; (iii) characterize locally the nature and anisotropy of the scatterers via their frequency response and radiation pattern.More generally, this work falls into a larger framework, which aims to develop a universal matrix approach that can be applied to any type of waves where multiple sensors can be used to shape incident wave-fronts and analyze reflected ones. This thesis describes this matrix approach in the ultrasound imaging context and paves the way towards a quantitative ultrasound imaging of soft tissues.
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, May 31, 2022 - 3:15:32 AM
Last modification on : Wednesday, June 1, 2022 - 3:37:36 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03682280, version 1


William Lambert. Matrix approach for ultrasound imaging and quantification. Acoustics [physics.class-ph]. Université Paris sciences et lettres, 2020. English. ⟨NNT : 2020UPSLS028⟩. ⟨tel-03682280⟩



Record views


Files downloads