Comparison of no-prior and soft-prior regularization in biomedical microwave imaging.

In the NIR case, the contrast and coupling effects (on a true inclusion) of the false inclusion are much more significant than in the corresponding microwave reconstructions where very modest influences were observed.[ 41 ] Additionally, the independence with which contrast in permittivity is recovered relative to conductivity; that is, the degree of estimated property parameter coupling is much less in the microwave case when spatial priors are invoked relative to the equivalent NIR problem where property parameter coupling is fairly significant (anywhere from 15% to 40% or more depending on noise level[ 28 38 ] compared to about 5% here).

also in 581 other papers