Leukocyte Segmentation in Giemsa-stained Image of Peripheral Blood Smears Based on Active Contour

Leukocytes are categorized into five groups in a peripheral blood microscopic image. A differential count of these various types of cells is used to determine the presence of an infection in the human body. Precise boundary of leukocyte and cytoplasm are necessary for extracting features of them. In fact texture, color, size and morphology of nucleus and cytoplasm make differences among different kinds of leucocytes. Otsu method has been used for thresholding in order to detect precise boundary of nucleuses in image and Active contour has been employed to find precise boundaries of Cytoplasm in it.Gradient Vector Flow of the thresholded image has been used as an external force for contour. After that desired features can be extracted in order to classify all types of leukocytes.