In the Proceedings of the 14th International Conference on Pattern Recognition, vol. 1, pp. 117-122, Brisbane, Australia, August 16-20, 1998.

Abstract

In this paper classification of surface defects is considered. The classification system consists of several classifiers whose outputs are combined in order to produce the final classification. The self-organizing maps (SOMs) are used as classifiers. Each SOM is taught unsupervised with examples of defects. Classification is based on the internal structure and the shape characteristics of defects. Texture features from the co-occurrence matrix and the gray level histogram are used to describe the internal structure. The set of simple shape descriptors is used for shape characterization The results of experiments with base paper defects are encouraging.