A morphological image processing method for locating myosin filaments in muscle electron micrographs

A morphological image processing based method for determining the positions of myosin filaments in electron micrographs of muscle cross-sections is described. The filaments are embedded in a noisy and variable background in the images, but lie on a relatively regular lattice. The filament positions are determined by a novel implementation of morphological grayscale reconstruction in which a threshold is optimised by using the lattice regularity of the derived filament markers. This approach leads to a robust algorithm. The method is applied to a number of muscle micrographs.