Multilevel Neural Network approach for general purpose quality diagnosis

In this paper we present a new approach for quality diagnosis of products whereas a neural network represents the main tool. Since flaws can be detected from the level of similarity between datasets, we describe how to compare two data partitions with a quite similar structure as frequently occurs in defect detection. The comparison is obtained by means of an unsupervised neural network algorithm and by the association of an undirected complete weighted graph structure. Then, a graph matching operation returns an estimation of the level of similarity between the datasets.