NEURO-FUZZY REASONING FOR OCCLUDED OBJECT RECOGNITION

To tackle the problem of occluded object recognition first we give a n ew interpretation to the multidimensional fuzzy reasoning and then rea lize that new interpretation through backpropagation-type neural netwo rk. At the learning stage of the neural network, fuzzy linguistic stat ements are used. Once learned, the nonfuzzy features of an occluded ob ject can be classified. At the time of classification of the nonfuzzy features of an occluded object we use the concept of fuzzy singleton. An effective approach to recognize an unknown scene which consists of a set of occluded objects is to detect a number of significant (local) features on the boundary of the unknown scene. Thus the major problem s fall into the selection of the appropriate set of features (local) f or representing the object in the training stage, as well as in the de tection of these features in the recognition process. The features sho uld be invariant to scale, orientation and minor distortions in bounda ry shape. The performance of the proposed scheme is tested through sev eral examples. (C) 1998 Elsevier Science B.V.