CLASSIFICATION OF SIMS IMAGES USING A KOHONEN NETWORK

The distributions of atomic and molecular ions of 12 different masses were measured by imaging SIMS at low mass resolution. Due to mass inte rferences, visual interpretation of the chemical phases represented in the different distributions is not possible. These single phases were extracted by classification using a Kohonen network. To demonstrate t his technique, the behavior of the Kohonen map is compared with manual classification. For determination of the optimal dimension of the net work (the number of nodes should be equal to the number of expected cl asses), and to reduce the artifacts due to noise and nonlinearities, p rincipal-component analysis was performed. Alternatively, the number o f necessary classes can be determined by a second classification of th e nodes of a Kohonen network that is sufficiently large with the help of dendrograms.