Very little work has been conducted on three-dimensional aspects of el
ectrical impedance tomography (EIT), partly due to the increased compu
tational complexity over the two-dimensional aspects of EIT. Neverthel
ess, extending EIT to three-dimensional data acquisition and image rec
onstruction may afford significant advantages such as an increase in t
he size of the independent data set and improved spatial resolution. H
owever, considerable challenges are associated with the software aspec
ts of three-dimensional EIT systems due to the requirement for accurat
e three-dimensional forward problem modelling and the derivation of th
ree-dimensional image reconstruction algorithms. This paper outlines t
he work performed to date to derive a three-dimensional image reconstr
uction algorithm for EIT based on the inversion of the sensitivity mat
rix approach for a finite right circular cylinder. A comparison in ter
ms of the singular-value spectra and the singular vectors between the
sensitivity matrices for a three-dimensional cylinder and a two-dimens
ional disc has been performed. This comparison shows that the three-di
mensional image reconstruction algorithm recruits more central informa
tion at lower condition numbers than the two-dimensional image reconst
ruction algorithm.