The May 1989 special issue of Optical Engineering is dedicated to a discussion of advanced optical infor mation processing techniques. The topics presented in this issue are intended to provide a discussion for present and future research developments in the field of optical information processing.

We present for the first time results of tests of a binary phase-only filter fabricated by standard VLSI techniques on a fused quartz substrate. We compare computer simulations with direct experimental optical bench tests. In general, the results agree well in terms of narrowness of the correlation peak and its height above the noise floor.

Recently, much attention has been paid to designing correlation filters that provide distortion-invariant pattern recognition. However, optical correlators can also be used to estimate the orientation and scale change in the object being observed. This paper presents two new methods for designing correlation filters that are aimed at determining the in-plane rotation and scale changes present in the input image (relative to a reference image). These methods assume that a known object is present in the input plane and that its rotation and scale are to be estimated. The first method is capable of determining both rotation and scale changes simultaneously. The second method assumes that the scale is fixed and estimates only the in-plane rotation. Both methods should prove useful in applications in which one needs to estimate the object orientation. Several simulation results are included to quantify the estimation accuracies and the computational complexities of the two methods.

This paper discusses a method of increasing the memory of an optical correlator through the use of matched filter arrays. The matched filters are sequentially addressed using an acousto-optic beam deflector.

Five types of discrete-valued correlation filters, including binary phase-only and ternary phase-amplitude types, were tested in simulations addressing specific character recognition problems. The filters were evaluated for classification accuracy and correlation efficiency. Transform-ratio filter ternary phase-amplitude filters provided the best classification performance but also the lowest correlation efficiency.

We use the kernel function of the circular Fourier and radial Mellin transforms to make spatial filters and extract shift, rotation, and scale invariant features from the 2-D correlation outputs. Using a small bank of filters, this approach allows pattern recognition for multiple input objects that is invariant to their independent distortions. These Fourier-Mellin spatial filters can be phase-only and binary phase-only filters and implemented in real time using commercially available spatial light modulators. These filters are suitable as preprocessors in artificial neuron networks for adaptive and self-organizing pattern recognition. Computer simulation results are shown.

Applications of morphological transformations on an optical symbolic substitution processor are presented. Simulation results of several morphological image processing operations (edge detection, hit-or-miss transformation, and skeletonization) are presented, and examples of their use in industrial inspection are given.

An approach to learning in a multilayer neural network is presented. The proposed network learns by creating interconnections between the input layer and the intermediate layer. In one of the new storage prescriptions proposed, interconnections are excitatory (positive) only and the weights depend on the stored patterns. In the intermediate layer each mother cell is responsible for one stored pattern. Mutually interconnected neurons in the intermediate layer perform a winner-take-all operation, taking into account correlations between stored vectors. The performance of networks using this interconnection prescription is compared with two previously proposed schemes, one using inhibitory connections at the output and one using all-or-nothing interconnections. The network can be used as a content-addressable memory or as a symbolic substitution system that yields an arbitrarily defined output for any input. The training of a model to perform Boolean logical operations is also described. Computer simulations using the network as an autoassociative content-addressable memory show the model to be efficient. Content-addressable associative memories and neural logic modules can be combined to perform logic operations on highly corrupted data.

A programmable optical associative memory for two-dimensional image retrieval is described. Both the stored images and the input image are displayed spatially; therefore, they can be updated in real time more conveniently. The integral product between the input image and the stored images is obtained by a nonlinear correlation technique that has a superior performance compared with conventional optical correlation techniques in the areas of light efficiency, correlation peak to sidelobe ratio, and correlation spot size. Thus, better quality images can be reconstructed, and the need for optical gain and optical feedback may be eliminated. The losses in the system are small since no holograms are employed.

A novel holographic associative memory that is able to identify individual words and insert word breaks into a concatenated word string is described. The system is shift invariant and has error correction capability. Experimental results obtained from the real-time system are demonstrated. A theoretical analysis of the system is presented, and practical limitations are discussed.

Although certain iterative optical processors promise algorithmic convergence at the speed of light, little attention is normally given to the consequences of different path lengths required within the processor on the processor performance. The resulting clock skew can have significant degrading effects on the predicted accuracy, stability, and speed of the processor. A similar problem occurs in iterative asynchronous artificial neural networks when, for example, the time delay between two neurons is proportional to their physical separation. In this paper, we show that in the absence of temporal dispersion, certain iterative algorithms have stable steady-state solutions that are independent of clock skew. Examples include stable linear feedback and feedback using soft (slowly varying) nonlinearities. Both are special cases of using a contractive operation in the feedback path. Such processing algorithms can have stable steady-state solutions that are independent of clock skew. Feedback using hard nonlinearities, on the other hand, can result in either an oscillatory or a steady-state solution that depends on the clock skew.

Experimental implementation of a holographic optical associative memory system using a thresholding microchannel spatial light modulator (MSLM) is presented. The first part of the system is basically a joint transform correlator, in which a liquid crystal light valve is used as a square-law converter for the inner product of the addressing and input memories. The MSLM is used as an active element to recall the associated data. If the device is properly thresholded, the system is capable of improving the quality of the output image.

The convergence mechanism of vectors in Hopfield's neural network in relation to recognition of partially known patterns is studied in terms of both inner products and Hamming distance. It has been shown that Hamming distance should not always be used in determining the convergence of vectors. Instead, inner product weighting coefficients play a more dominant role in certain data representations for determining the convergence mechanism. A trinary neuron representation for associative memory is found to be more effective for associative recall. Applications of the trinary associative memory to reconstruct machine part images that are partially missing are demonstrated by means of computer simulation as examples of the usefulness of this approach.

A darkroom exposure meter has been designed to compute from a single measurement sequence both the exposure time and the paper contrast necessary to accept the full tonal scale of the negative. The photographer moves a phototransistor probe across the projected image to sample all areas that are to show gradation in the finished print. A microcomputer accepts the digitized probe output, compares the whole range of readings with paper characteristics stored in its memory, and calculates contrast grade and exposure time. A prototype constructed on this basis has demonstrated the feasibility of this concept. Such a device can greatly improve darkroom productivity.

By extracting the basic rectangular features of an object from a digitized Fourier transform image, one can calculate the scale factor and rotational angle of a rectangular object and use them to derotate and scale the input image. Thus, a scale- and rotational-invariant optical correlator is achieved. The accuracy of the computed scale and orientation parameters yielded by this approach ensures nominal correlator performance for object recognition and tracking applications. The concept and system implementation of such an intelligent optical correlator are discussed, and the results of performance analysis are presented.

Comparative holographic moire interferometry (CHMI) permits defect detection by comparison of the mechanical responses of two macroscopically identical specimens at equal loading. In the present work a technique of moire fringe separation is suggested that is amenable to automatic readout. Two approaches are proposed and experimentally verified: phase shifting of the carrier interference pattern and auxiliary carrier fringe generation with subse-quent optical filtering. The phase shifting method is applicable to an additive moire scheme with two reference beams. Auxiliary carrier fringe generation can be used in both additive and multiplicative CHMI. The frequency constraints of the latter method due to the optical recording arrangement and mechanical responses of the tested specimen are analyzed.

We develop a progressive image transmission technique based on a prediction/residual encoding approach in which the predictive and residual image components are both progressively encoded. The predictive component is taken to be an interpolative representation of the image and is progressively transmitted first. The goal of this first stage is to transmit a "recognizable" image with the fewest possible data so that the user has the opportunity to terminate the transmission if the image is not the desired one. However, if the user decides to retain the image, the second stage follows, in which the residual component is progressively transmitted to refine the image quality. Specifically, the Knowlton hierarchy is used to transmit the predictive component, and a tree-structured vector quantizer hierarchy is used to transmit the residual component. Subject terms: digital image coding; progressive transmission; hierarchical coding structures; vector quantization.

A simple theoretical model is presented that allows calculation of the image produced by a spherical absorbing particle illuminated by monochromatic, coherent laser light. Results presented in this paper are restricted to a single-lens imaging system, although generalization to more complex imaging system configurations would be straightforward. The method uses classic Lorenz-Mie scattering theory to obtain the electro-magnetic field external to an absorbing spherical particle and a Fourier optics approach to calculate the intensities in the image plane. Experimental results evaluating focus characteristics are examined for 50 um diameter water droplets using an N2 laser imaging system in conjunction with a digital image processor, and the experimental images are compared to the results of the theoretical model. Comparative focus criteria results are particularly useful in aerosol science research involving dynamic particle size measurements in which criteria for focus and depth of field must be established.

I haven't presented any publication statistics for the journal for quite some time, so I decided to bring you up to date just a bit. The figures shown below are actual figures for each of the categories listed for the three-year period 1986-1988.

This book is a compilation of research papers written by R. V. Jones in the areas of measurement and instrument design, with added commentary. A unifying theme of the ppapers is the concept of precision in measurement and instrument design. Fifty years of research, from prior to the Second World War to the present, are covered.