The scanning electron microscope produces nanometer-resolution surface images of biological samples preserved in a life-like state. Extracting three-dimensional information from these two-dimensional images has been the subject of long and ongoing research. We present here a general method and theoretical basis for reconstructing the surfaces of SEM specimens imaged from multiple directions by back-projection. The resulting reconstructions are faithful representations of the original specimen geometry, even when the input images are blurred and have low signal-to-noise ratio.
Copyright 2009 Elsevier B.V. All rights reserved.