Neural stem cell segmentation using local complex phase information

Segmentation of neural stem cells is the preliminary step to treat and cure several brain neural diseases. There exist a number of methods to accomplish this task. However, all of these methods suffer from some problems, such as high intensity variation sensitivity, human interaction and high computational complexity. In this paper we proposed a novel edge-detection-based neural stem cell image segmentation algorithm using the local complex phase characteristics. The proposed method is an illumination and contrast invariant measurement of edge significance. Our contributions are that, local weighting summation Gaussian kernel convolution and a new model for phase deviation weighting function are introduced into the proposed model to improve the local phase measurement. In experiments, we show that the proposed method is more accurate and reliable than three existing gradient-based edge detection algorithms and Kovesi's model for neural stem cell image segmentation.