Multichannel restoration with limited a priori information[electron microscopy of biological macromolecules]
Vrhel, M.J.; Unser, M.
Image Processing, IEEE Transactions on
Volume 8, Issue 4, Apr 1999 Page(s):527 - 536
Digital Object Identifier 10.1109/83.753740
Summary:We introduce a method for multichannel restoration of images in
which there is severely limited knowledge about the undegraded signal,
and possibly the noise. We assume that we know the channel degradations
and that there will be a significant noise reduction in a postprocessing
stage in which multiple realizations are combined. This post-restoration
noise reduction is often performed when working with micrographs of
biological macromolecules. The restoration filters are designed to
enforce a projection constraint upon the entire system. This projection
constraint results in a system that provides an oblique projection of
the input signal into the subspace defined by the reconstruction device
in a direction orthogonal to a space defined by the channel degradations
and the restoration filters. The approach achieves noise reduction
without distorting the signal by exploiting the redundancy of the
measurements
View citation and abstract |
Cart
