Fuzzy patch label relaxation in bone marrow cell segmentation

Performing differential white blood cell counts in a bone marrow preparation is a crucial step in diagnosing various disease states. It is a tedious task to locate, identify, and count these classes of cells. In an active project, we are investigating the automation of this task. In this paper, we present a cell segmentation approach which utilizes the principle of least commitment. We use the watershed algorithm to perform an "oversegmentation" of the image where each primitive patch is no bigger than one of the cell components (nucleus, cytoplasm, red blood cell, or background). We assign memberships to these patches and "relax" the patch label memberships in order to obtain more consistent labels for merging into cell objects.