Automatic segmentation of neurons from fluorescent microscopy imaging

Abstract

Automatic detection and segmentation of neurons from microscopy acquisition is essential for statistically characterizing neuron morphology that can be related to their functional role. In this paper, we propose a combined pipeline that starts from the automatic detection of the soma through a new multiscale blob enhancement filtering. Then, a precise segmentation of the detected cell body is obtained by an active contour approach. The resulted segmentation is used as initial seed for the second part of the approach that proposes a dendrite arborization tracing method.