A shape analysis method to detect dendritic spine in 3D optical microscopy image

Dendritic spines play an important role in synaptic transmission. Spines are small protrusions formed on the shafts of dendrites and they are manifest of many neurological conditions. Three dimensional images of dendrites and spines are obtained by confocal and two-photon laser scanning microscopy. We present an image processing method based on shape analysis to detect spines and separate spines from the dendrites for quantitative study. The method applies two grassfire procedures to first find the tips of spines and then reversely locate the boundary between spines and dendrites. The boundary is determined as the propagation front that undergoes the maximum change in plane area of the grassfire procedure. We use real data to demonstrate the performance of the method.