Automated Recursive Segmentation of Large Neocortical Images Using Standard Deviation as Termination Criteria

In this paper we present an improved segmentation algorithm that recursively explores various thresholding levels until it reaches a termination criteria. This segmentation algorithm is based on earlier work adapting Otsu's thresholding approach to myelinated bundles of axons in cortical tissue. Experimentation using over 120 images has confirmed that this termination criteria provides visibly acceptable segmentation in an automated fashion