Zur Anwendung der statistischen Formanalyse

In the following work we propose an introduction to Statistical Shape Analysis and adapt the procedures to three and four dimensional objects. Especially, in the case of four dimensional objects we show one possible approach of Dynamical Shape Analysis. In Shape Analysis we use parameters of the distribution as the mean shape, a representative shape of a group, to test the possibility of discriminating different kind of groups of objects. First-time we are using in the concept of Shape Analysis the variance. Furthermore we combine the Neural Networks with the Shape Analysis and also the tests used for discriminating groups with Answer Tree. To classify the objects well known procedures as Logistic Regression, Neural Networks, Discriminant Analysis and Answer Tree are used. For the application more than one hundred renal tumors in childhood, also more than one hundred electronic nose sensor data for different kind of odor quality and concentration as well as the behaviour of around fifty sexual serial murderer are measured and used. In all the disciplines a new approach in measurement of data is shown. The work shows furthermore, that all three fields Shape Analysis is applicable. The results of Shape Analysis and the classification afford the discrimination of different kind of objects. Shape Analysis could be one answer for classifying image data in the light of increasing image data.