Pattern recognition of nucleated cells from the peripheral blood

摘要

A data set consisting of 2259 nucleated cells from the peripheral blood, including immature cells, was systematically investigated using image processing and pattern recognition techniques. A total of 94.2% of all cells could be correctly classified into one of 9 possible main classes. Several subsets were also investigated. In particular, the cell types in the neutrophilic maturation sequence were extensively explored.