Automatic Recognition of Five Types of White Blood Cells in Peripheral Blood

Abstract

AbstractAn automatic system which is capable of recognizing white blood cells can assist hematologists in the diagnosis of many diseases. In this paper, we propose a new system based on image processing techniques in order to recognize five types of white blood cells in the peripheral blood. To segment nucleus and cytoplasm, a Gram-Schmidt orthogonalization method and a snake algorithm are applied, respectively. Moreover, three kinds of features are extracted from the segmented areas and two groups of textural features extracted by Local Binary Pattern (LBP) and co-occurrence matrix are evaluated. Best features are selected using a Sequential Forward Selection (SFS) algorithm and performances of two classifiers, ANN and SVM, are compared. In this application, the best result is obtained using LBP as the textural feature and SVM as the classifier. In sum, the results demonstrate that the methods are accurate and fast enough to execute in hematological laboratories.KeywordsWhite blood cellperipheral bloodsegmentationtextural featurefeature selectionclassification