Geometry-Independent Speed of Sound Reconstruction for 3D USCT Using Apriori Information

We develop a 3D Ultrasound Computer Tomograph (3D USCT) for early breast cancer detection. It is capable of transmission and reflection imaging. Transmission tomography provides speed of sound as well as attenuation images and serves as accurate background medium estimation for reflection tomography. We use Algebraic Reconstruction Technique (ART) algorithms for transmission tomography. The main algorithmic step in ART algorithms is the solution of a large system of equations. There are serveral solvers available, however erroneous and noisy data pose problems for the solving process. Our goal is to include apriori knowledge in the reconstruction to obtain better image quality and less artifacts due to noise. By including transducer directivity, domain-specific interpolation and other apriori information we improved the reconstruction images and decreased noise artifacts. We show exemplary speed of sound reconstructions of a breast phantom.