Teager energy based blood cell segmentation

A new method for fast and simple segmentation of peripheral blood smears using the Teager energy operator (TEO) is presented. The two-part algorithm efficiently segments leucocytes present in the smear into nucleus and cytoplasm. The local mean weighted high pass filtering property of the Teager energy operator is used to identify and distinguish the nucleus of white blood cells. Cytoplasm present in the white blood cells is segmented using selective mathematical morphology. Experimental results are shown for a blood smear containing multiple nuclei. It was observed that modified histogram developed using the TEO can efficiently segment even multiple nuclei. Experiments have been conducted to verify the accuracy of the proposed segmentation scheme for all the major types of leucocytes. Simulation results have shown that our algorithm segments nuclei effectively even at low percentage of impulse noise.