Tensors of diffusion, deformation (stress and strain), and conductivity are
physical quantities of biological tissue, which can be used to, characteri
ze particular disease states. Tensor tomography may be a useful imaging tec
hnique for eliciting these tensor quantities when applied in conjunction wi
th an imaging modality such as magnetic resonance imaging (MRI). This paper
presents a method for reconstructing a 2x2 second-order tensor field from
scalar projection measurements of the tensor field. The reconstruction of t
he four components of a 2x2 tensor may require as many as four distinct sca
lar measurements for each projection ray through the tensor field. Fourier
projection theorems have been developed for the reconstruction of a tensor
field which is decomposed into solenoidal and irrotational components. Resu
lts of a computer simulation that demonstrate the validity of the mathemati
cal formulations are presented. A method is also proposed to obtain a diffu
sion tensor field tomographically from MRI projection measurements of the d
iffusion tensor field. The method is evaluated experimentally and results o
f an MRI phantom study are presented.