INVARIANT REPRESENTATION AND HIERARCHICAL NETWORK FOR INSPECTION OF NUTS FROM X-RAY IMAGES

An X-ray based system for the inspection of pistachio nuts and wheat k ernels for internal insect infestation is presented. The novelty of th is system is twofold. First, we construct an invariant representation of infested nuts from X-ray images that is rich, robust, and compact. Insect infestation creates a tunnel, in the X-ray image, with reduced density of the natural material. The tunneling effect is encoded by li nking troughs on the image and constructing a joint curvature-proximit y distribution table for each nut. The latter step is designed to acce ntuate separation of those tunneling effects that are due to the natur al structure of the nut. Second, since the representation is sparse, w e partition the joint distribution table into several regions, where e ach region is used independently to train a backpropagation (BP) netwo rk. The outputs of these subnets are then collectively trained with an other BP network. We show that the resulting hierarchical network has the advantage of reduced dimensionality while maintaining a performanc e similar to the standard BP network. (C) 1996 John Wiley & Sons, Ind.