This paper describes an automated classification of the leukocyte into 4 categories (Lymphocyte, Segmented Neutrophil, Band Neutrophil, Eosinphil). The following 2 processings were performed by software in order to extract the features of a leukocyte. 1. the elimination of the red cells 2. the separation of a leukocyte into nucleus and cytoplasm The features were measures of such visual concepts as nucleus area, nucleus shape, nucleus intensity, cytoplasm area, cytoplasm intensity, and cell perimeter. A minimum distance classifier was designed by a training set of 42 cells. Tests on an independent testing set of 27 cells showed classification errors on the order of 26 percent.