A formulation for 2-D array neural nets for electro-optical implementation

Summary form only given, as follows. A formulation is described for 2-D neural nets which is suitable for the processing of 2-D array images. The formulation essentially processes a 2-D input array to be pre- and postmultiplied by a matrix of weights. Then the resulting 2-D array of data is processed via the usual sigmoidal thresholding and feedback which are characteristic of neural nets. A description is also given of two electrooptical implementations of the formulation: (i) it includes adaptations of the available incoherent optical techniques for the real-time multiplication of three matrices, and (ii) it includes adaptations of vector-matrix multiplication and the sum of the outer product of vectors. The second requires reduced optical hardware for its implementation.<>