Sharpening of neurite morphology using complex coherence enhanced diffusion

The study of the molecular mechanisms involved in neurite outgrowth and differentiation, requires essential accurate and reproducible segmentation and quantification of neuronal processes. The common method used in this study is to detect and trace individual neurites, i.e. neurite tracing. The challenge comes mainly from the morphological problem in which these images contain ambiguities such as neurites discontinuities and intensity differences. In our work, we encounter a bigger challenge as the neurites in our images have a higher density of neurites. In this paper, we present a hybrid complex coherence-enhanced method for sharpening the morphology of neurons from such images. Coherence-enhanced diffusion (CED) is used to enhance the flowlike structures of the neurites, while the imaginary part of the complex nonlinear diffusion of the image cancels the appearance of 'clouds'. We also describe an elementary method for estimating the density of neuritis based on the obtained images. Our preliminary results show that the proposed methodology is a step ahead toward an effective neuronal morphology algorithm.