Generating three-dimensional neural cells based on bayes rules and interpolation with thin plate splines

Abstract

In this paper the use of Bayes rules and interpolation functions is proposed in order to generate three-dimensional artificial neural cells incorporating realistic biological neural shapes. A conditional vectorial stochastic grammar has been developed to control and facilitate the parallel growth of branching. A L-parser (parser to L-Systems) has also been developed to guarantee that the grammar is free from mistakes before its use. This parser has also the function to generate a group of points corresponding to the morphologic structure of a neural cell. These points are visualized in a three-dimensional viewer especially developed to show the neural cell generated. © Springer-Verlag Berlin Heidelberg 2003.