The dependence of EIT images on the assumed initial conductivity distribution: a study of pelvic imaging

Abstract

EIT measurements on humans are often made in regions of the body where the conductivity distribution is far from uniform. This paper addresses the problem of deriving accurate quantitative data in one such region: the conductivity changes associated with the accumulation of blood in the pelvic bowl. A computer map of the bone in the pelvic region was constructed, from which an appropriate reconstruction matrix was generated. Both computer simulations and tank tests were performed to assess whether this bone reconstruction matrix produced impedance images with closer fidelity to the measured object than images produced using a reconstruction based on a uniform conductivity distribution. As expected, images produced by the computer simulation indicated that the bone reconstruction matrix produced images of better fidelity than did the uniform reconstruction matrix. However, in the case of the tank data only a moderate improvement was achieved. The reconstruction matrix based on a uniform conductivity distribution was found to produce satisfactory images for both bone and near-uniform objects, but for regions further into the pelvic bowl, where the signal was lower, the uniform reconstruction matrix was less satisfactory.

A Robust Current Pattern for the Detection of Intraventricular Hemorrhage in Neonates Using Electrical Impedance Tomography

“…Earlier studies1,3,16 showed that incorporating a more accurate forward model that matches the conductivity structures of the imaged object can improve image quality in general. For example, inclusion in forward models of a bone model,16 a four-shell head model,3 or including anisotropy1 all demonstrated improvements in image quality, image resolution, and localization errors.…”

“…For example, inclusion in forward models of a bone model,16 a four-shell head model,3 or including anisotropy1 all demonstrated improvements in image quality, image resolution, and localization errors. However, realistic models need to be used carefully, since with inaccurate prior information, they may yield images worse than those reconstructed with a simple model 14.…”

“…In Bagshaw et al ,3 it was noted that using a full-shelled head model produced worse images than using the homogeneous head model when reconstructing the head phantom data, because the resistivity ratio between the skull and brain region for the tank was too dissimilar to the value assumed in the model. Meeson et al 16 also concluded that using a uniform matrix seems less likely to generate artifacts when the underlying conductivity distribution is unknown. Bagshaw et al 3 also studied the effect of using RG in the forward model.…”

“…Initially, forward models and sensitivity matrices may be formulated using a simple model such as a sphere containing a homogeneous conductivity distribution. In principle, better results from the inverse solution should be obtained if the forward model (the prior distribution) exactly matches the imaged object in terms of internal conductivity distribution and external geometry 1,3,16. However, it is difficult to obtain an accurate model for an individual head because of the difficulty of determining the locations of electrically relevant structures such as scalp, skull, CSF, grey, and white matter from modalities such as MRI and CT.…”

“…Another problem that has been frequently addressed in the EIT literature 1,3,14–16 is the benefit of incorporating an accurate reconstruction model to improve the image quality. In this study, we also address the problem of whether the use of a simple homogeneous spherical might be sufficient for the purpose of quantification, and compare image and quantification results with and without realistic reconstruction modeling.…”