SEQUENTIAL PEAK DETECTION FOR FLOW CYTOMETRY

Circulating tumor cells in blood are identified by means of sequential peak detection taking into account the memory and real time applicability constraints. Three different spatial domain algorithms: derivative approach, energy detector and baseline method are compared with three different peak detection algorithms based on machine learning: linear and nonlinear support vector machines and artificial neural networks. Performance of the peak detection algorithms are tested on both synthetic and real data. Experimental results indicate superiority of machine learning algorithms over the other three algorithms which are widely used in practice. Due to Gaussianity assumption in the signal model, a linear support vector machine is found to be as good as other machine learning schemes.