Multi-thresholding for neural stem cells image segmentation based on two clustering methods

Automatic segmentation and tracking systems are expected to be useful tools for monitoring the proliferation and differentiation of neural stem cells.This study compared the self-organizing map(SOM) based and k-means clustering based multi-thresholding method for segmenting neural stem cell images.The results indicate that these two clustering methods both effectively construct markers for the watershed method,with the foreground markers acquired by SOM having fewer contaminations than the k-means makers.