Automatic Morphological Evaluation of Endothelial Cells Using Different Classification Methods

The cell morphology analysis is applicable to pathophysiology studies in biological samples. In this work, digital images of Human Umbilical Vascular Endothelial Cells (HUVEC) were classified according to their morphological properties, to help the detection of functional and/or structural anomalies for the study of angiogenesis, a process by which new capillaries are formed from pre-existing capillaries. The automatic classification was produced by the algorithms: support vector machine (SVM), k-Nearest Neighbors (k-NN), and decision trees (DT), with three classes: circular, elongated deformed (elongated), and slightly elongated deformed (others deformations). The processes of cell migration and proliferation could be correlated with this classification. The sensitivity values for all three methods exceed 95%. The highest accuracy value, 98.89%, was reached by SVM method. Results shows that it is feasible to use these three methods for the classification of HUVEC.