A comparison between neural networks and k-nearest neighbours for blood cells taxonomy

Constitutive properties of living cells are able to withstand physiological environment as well as mechanical stimuli occurring within and outside the body. Any deviation from these properties would undermine the physical integrity of the cells as well as their biological functions. Thus, a quantitative study in single cell mechanics needs to be conducted. In this paper we will examine fluid flow and Neo-Hookean deformation related to the rolling effect. A mechanical model to describe the cellular adhesion with detachment is here proposed. We develop a first finite element method (FEM) analysis, simulating blood cells attached on vessel wall. Restricting the interest on the contact surface and elaborating again the computational results, we develop an equivalent spring model. Our opinion is that the simulation notices deformation inhomogeneities, i.e., areas with different concentrations having different deformation values. This important observation should be connected with a specific form of the stored energy deformation. In this case, it loses the standard convexity to show a non-monotone deformation law. Consequently, we have more minima and the variational problem seems more difficult. Several numerical simulations have been carried out, involving a number of human cells with different mechanical properties. All the collected data have been subsequently used to train and test suitable soft computing models in order to classify the kind of cell. Obtained results assure good performances (4.7% of classification error) of the implemented classifier, with very interesting applications. © 2010 Springer-Verlag.