Automated tracing of microglia using multilevel thresholding and minimum spanning trees

Abstract

Microglia are immune cells of the central nervous system. Good knowledge about their morphology leads us to a better understanding of their functionality. In this article we propose an automated method to trace microglia in microscopy images. Our approach is powered by two main algorithms: multilevel thresholding (MT) and minimum spanning tree (MST). MT quantizes intensities of image pixels to several levels, then we sample each level with different rates to produce seed points. Each seed point belongs to a level and has a specific value, therefore they can be prioritized. The tree structure of microglia is known a priori, so we apply the MST to the prioritized seed points to preserve this feature. Results show that our proposed method is fast and accurate in reconstruction of large images.