Extracting useful information from three dimensional nuclear magnetic resonance (NMR) images usually requires some form of explicit information about the object one is trying to visualise. This knowledge can either be in the form of a computer based identification method, such as edge identification, or can rely on human inspection, of either two dimensional slices, or three dimensional images stripped back with the aid of paint packages. The method considered represents an attempt to construct an automated process for the identification of such objects by using a genetic algorithm (GA). The method has proved successful and has been tested on several problems including the location and visualisation of three dimensional objects from three dimensional images of animal tissue.
展开▼