Automatically Extracting Cellular Structures from Images Generated via Electron Microscopy

In this paper, we consider mathematical techniques for locating cellular structures in digital images generated via electron microscopy. We approach this problem in two steps: a pre-processing denoising stage and a segmentation stage. For image denoising, we will limit our discussion to Partial Differential Equation (PDE) based methods, primarily focusing on diffusion and total variation methods. Segmentation will also be accomplished via PDE methods. Combining these two stages, we can automatically extract some cellular structures from images produced by electron microscopes. We will then look at the parallelization of these techniques for computing on a cluster computer environment and provide a basic analysis of the performance of the parallel code.