An optical matrix-vector multiplier has ben efficiently used for photonic implementation of Hopfield network model, which is used for binary pattern recognition. Training matrices are recorded on electrically addressed spatial light modulator, where each matrix is composed of the same row of each pattern, that the network is being trained with. After training, if an unknown pattern is presented to the network in the form of a vector, the output vector is obtained by the element that has the highest magnitude through a winner- take-all algorithm. Pattern can be recognized even if the input is noisy and distorted.
An improved optical matrix-vector multiplication is performed by convolution process. The multiplicated binary numbers are represented by on/off states of light sources and the multiplier binary numbers are recorded on a spatial light modulator. Cylindrical optics is used as free space interconnection. The convolution coefficients are recorded on a CCD array. The output of the CCD array are added in a computer to yield the result of multiplication. The operation is completely digital and needs no analog to digital conversion. Because of parallel operation in two dimensions, the processing speed is greatly increased.
The main difficulty in the realization of intelligent processing and computing techniques lies with the requirements of enourmous amount of machine computations and therefore must operate in a parallel processing environment. In this context interference-free parallel processing capability of optics technology can be efficiently exploited. Optical computing system can interconnect and operate on two dimensional input-output data structures and can store data in three dimension. Therefore the implementation of neural network models in optical domain is likely to be more effifient. The optical neural computing techniques with facilities of intelligent processing have become versatile tools for pattern identification, classification and recognition. The paper introduces a representative optical neural computing system developed for intelligent pattern association tasks. The system can associate patterns from noisy and incomplete informations. The space invariant properties can also be introduced during identification of patterns.
Realization of a two-dimensional discrete Walsh transform (DWT) by matrix-vector multiplication in an optical architecture is presented. Although data are entered sequentially, the parallelism of optical architecture is exploited in the proposed system and, therefore, the computation time required to obtain the DWT is less when compared to its equivalent electronic system. The hardware realization is achieved with the help of a microprocessor.
The paper suggests a method of computation of Discrete Walsh Transform (DWT) by matrix vector multiplication which is easier to perform in 2D optical system where processing can be done parallely in 2D. The computation of DWT in such an optical computing architecture is much easier to perform and requires less processing time. A methodology is also proposed where a microprocessor is utilized to sequentially process matrixvector operations on digitized data. The result is stored in the memory for display and/or read out. I.