Segmentation of neuron and measurement of optically programed neurite growth: Fast automation via Bayesian thresholding

Abstract

The variability and complex dynamics of cell morphology make the automated segmentation of neurons in microscopic images a rather difficult task. To fully leverage modern computational power in large-scale analysis of such biological images, automation is necessary. In this paper, we present an automated approach to segmenting individual cells from their surroundings, and test it on time-lapse images of hipppocampal neurons during neurite initiation and extension. Noting that active contour based methods are usually accurate, but computationally expensive and slow, we propose a fast hybrid approach that combines Chan-Vese active contour segmentation with Bayesian thresholding for segmentation of neuron and measurement of neurite growth dynamics. Our approach demonstrated upto two-hundred-fold faster quantification of growth dynamics compared to the pure Chan-Vese segmentation.