This special issue on Visual Communications and Image Processing contains 14 papers that cover a wide spectrum in this fast growing area. For the past few decades, researchers and scientists have devoted their efforts to these fields. Through this long-lasting devotion, we witness today the growing popularity of low-bit-rate video as a convenient tool for visual communication. We also see the integration of high-quality video into broadband digital networks. Today, with more sophisticated processing, clearer and sharper pictures are being restored from blurring and noise. Also, thanks to the advances in digital image processing, even a PC-based system can be built to recognize highly complicated Chinese characters at the speed of 300 characters per minute. This special issue can be viewed as a milestone of visual communications and image processing on its journey to eternity. It presents some overviews on advanced topics as well as some new development in specific subjects.

Let me begin by thanking Dr. Hsing for inviting me to speak to you today. It is a rare event, at least in my experience, to have the opportunity to tell a captive audience of the leaders in one's field exactly what one thinks about the work that has been done up to this point. History is always worth studying if for no other reason than to find some clues to the future. If we carefully examine which approaches had good results and which did not, then we may be able to direct our work more effectively.

The transmission of video information through a narrowband channel requires a significant degree of image compression. In addition to the actual coding procedure, the frame rate is reduced, the spatial sample rate is lowered, and the color components are further filtered. Through a review of a practical and flexible image compression configuration, the important technical aspects of video coding are presented. The complexity of the currently developed system is discussed for operation over a 64 kbit/s communication channel. The specific and primary considerations include the following: After a reduction of temporal and spatial resolution, the luminance and color difference signals are divided in pixel blocks of 16 x 16 and 8 x 8, respectively. Then they are passed on to a motion-compensated differential pulse-code modulation coding system with a subsequent discrete cosine transform of the prediction error. The appropriate quantization in performed by means of several Huffman tables.

Pyramid data structures have found an important role in progressive image transmission. In these data structures, the image is hierarchically represented, with each level corresponding to a reduced-resolution approximation. To achieve progressive image transmission, the pyramid is transmitted starting from the top level. However, in the usual pyramid data structures, extra significant bits may be required to accurately record the node values, the number of data to be transmitted may be expanded, and the node values may be highly correlated. In this paper, we introduce a reduced-difference pyramid data structure in which the number of nodes, corresponding to a set of decorrelated difference values, is exactly equal to the number of pixels. Experimental results demonstrate that the reduced-difference pyramid results in lossless progressive image transmission with some degree of compression. By use of an appropriate interpolation method, reasonable quality approximations are achieved at a bit rate less than 0.1 bit/pixel and excellent quality approximations at a bit rate of about 1.3 bits/pixel.

An interframe hierarchical vector quantizer (IHVQ) is presented that is capable of encoding image sequence scenes at rates below 0.3 bit per pixel per frame. A regular decomposition quadtree method is used to segment the interframe differential signal into homogeneous regions of different block size. Small blocks representing high contrast moving boundaries (i.e., the impulsive component of the difference signal) are vector quantized, whereas large blocks typically representing smooth regions of the image are encoded by the local sample mean of the region. The IHVQ system has a signal-to-noise ratio (SNR) at least 2 dB higher than the SNR of the interframe mean-reconstructed quadtree coding system.

Fractals and related mathematical models of chaotic phenomena have become active areas in research and applications in many diverse fields. These include weather forecasting, structural analysis, mapping, entertainment, biology, and many others. Aspects of these fields appear to be a significant part of an emerging inter-disciplinary science of complexity and complex systems in which significant understanding can be attained through the analysis of "simple" nonlinear interactions of large numbers of primitive simple objects. The uses of these models in the rapidly developing field of image computing (i.e., the digital computing aspects of the image sciences: image, analysis, communications, and synthesis) are leading to many innovations. This paper describes a range of activities illustrating significant new developments and a trend toward unification in image computing using these techniques.

This tutorial paper discusses the use of successive-approximation-based iterative restoration algorithms for the removal of linear blurs and noise from images. Iterative algorithms are particularly attractive for this application because they allow for the incorporation of prior knowledge about the class of feasible solutions, because they can be used to remove nonstationary blurs, and because they are fairly robust with respect to errors in the approximation of the blurring operator. Regularization is introduced as a means for preventing the excessive noise magnification that is typically associated with ill-posed inverse problems such as the deblurring problem. Iterative algorithms with higher convergence rates and a multistep iterative algorithm are also discussed. A number of examples are presented.

Nonlinear filters are used in many applications, including speech and image processing, owing to their ability to suppress noise and preserve signal features such as edges. This study presents a performance evaluation of non-linear filters derived from the robust point estimation theory. The first part of the work is a classification of various approaches to nonlinear filtering into three types of estimators according to the process of the filter. The second part is a computer implementation and evaluation of all of the filters discussed. Finally, a summary of experimental results is presented.

This paper presents the application of the regularizing operator (RO) approach in discrete image restoration. This is accomplished by extending the applicability of the RO approach to discrete inverse problems. The concept of the RO is initially presented, and the necessary condition for any operator to be considered as an RO is proposed. On the basis of this condition, a complete mathematical framework for the formation of specific ROs is developed. It is then shown that several operators already known in the literature, including those based on constrained least squares and parametric projection filters, can be considered as special applications of the concept of the RO. On the basis of a new optimality criterion introduced for image restoration, a new class of ROs is then proposed. Several computational issues related to the implementation of an image restoration system are then considered. Under certain assumptions regarding the image formation process, a number of exact regularizing algorithms are proposed. Finally, experimental results are presented, compared, and discussed.

A prototype point-to-point packet transmission system carrying a DS-3 rate, NTSC-quality digital video signal has been made and successfully demonstrated. Fixed bit-rate video signals are transported in packet form using the asynchronous transfer mode (ATM) technique. A packet assembler integrates the transmit terminals' information into an ATM data stream suitable for application to a broadband ISDN. A packet disassembler distributes the incoming data to the proper receive terminals. The flexible and modular architecture of both the packet assembler and disassembler allows service equipment to be physically remote and provides easy expansion or reconfiguration of the service channels in the future. Service clock information not carried by the network is recovered at the receive end by use of a digital phase-locked loop.

Integration of images from different sensing modalities can produce information that cannot be obtained by viewing the sensor outputs separately and consecutively. This paper introduces a hierarchical image merging scheme based on a multiresolution contrast decomposition (the ratio of a low-pass pyramid). The composite images produced by this scheme preserve those details from the input images that are most relevant to visual perception. The method is tested by merging parallel registered thermal and visual images. The results show that the fused images present a more detailed representation of the depicted scene. Detection, recognition, and search tasks may therefore benefit from this new image representation.

An intelligent optical Chinese character recognition system using accumulated stroke features has been developed to solve the input problem of Chinese characters. The hardware architecture of the system is built on an IBM PC-AT with three extension boards: the preprocessor board, the feature extraction board, and the matching recognition board. The system can recognize, at the same time in the same program, either printed or handwritten Chinese characters of different styles and sizes. At present, a total of 5401 commonly used Chinese characters can be recognized. Results show that 99% of printed characters and 90% of constrained handwritten characters can be correctly recognized, at a speed of about 300 characters per minute.

One of the most pervasive problems in the scholarship of classical paintings is that of authenticity. Traditionally, the attribution of a work of art rests on the subjective opinion of an art historian bolstered by scientific data pertaining to the types and possibly the ages of the materials of the artwork. To expand the range of technical information that may be applied to the painting authentication problem, the methods of computer image processing (IP) have been employed to compare the techniques in two paintings. One is the Mona Lisa del Gioconda by Leonardo da Vinci. The other is known as the Isleworth Mona Lisa and has also been attributed to Leonardo by a few scholars. Computer IP was used to compare statistical and geometrical features of the two paintings. It emerged that the Isleworth work is not a copy of the Louvre painting but does have numerous similarities in composition and execution. These findings lend support to the theory that the Louvre Mona Lisa may be a portrait of Costanza by Leonardo that had been thought lost.

Unacceptable moire distortion may result when images that include periodic structures such as halftone dots are scanned. In the frequency domain, moire patterns correspond to visible aliased frequencies. In the spatial domain, moire patterns are evident as cyclic changes in the size of halftone dots, producing visible periodic "beat" patterns. Moire pattern formation depends on the following factors: (1) the halftone screen frequency, (2) the scan frequency, (3) the angle between the scan direction and the halftone screen, (4) the scanner aperture size and shape, (5) quantization errors from the thresholding operation, (6) scanner and printer noise, and (7) the ink flow in the paper during printing. This paper analyzes the visibility of moire patterns in terms of these factors. In addition, the paper describes an approach to reducing the visibility of moire patterns by directly manipulating the moire formation factors. With an appropriate selection of the scan frequency and screen angle for a given screened image, moire beat frequencies in the scanned image can be reduced to a subvisible level. This approach thereby achieves moire reduction without the need for scanning at extreme scan frequencies or postscan image processing. The strengths of this ap-proach are (1) no data volume increase due to high frequency scan and (2) no need for time-consuming postscan processing. Computer-simulated and actual scan images are presented to illustrate the approach.

A preliminary image quality measure that takes into account two major sensitivities of the human visual system (HVS) is described. The sensitivities considered are background illumination level and spatial frequency sensitivities. Given a digitized monochrome image, the algorithm produces, among some other figures of merit, a plot of the information content (IC) versus the resolution in units of pixels. The IC is defined here as the sum of the weighted spectral components at an arbitrary specified resolution. The HVS normalization is done by first intensity remapping the image by a monotonically increasing function representing the background illumination level sensitivity, followed by a spectral filtering to compensate for the spatial frequency sensitivity. The developed quality measure is conveniently parameterized and interactive. It allows experimentation with numerous parameters of the HVS model to determine the optimum set for which the highest correlation with subjective evaluations can be achieved. The preliminary results are promising.

A theoretical scheme to describe the infrared contrast field left by a surface ship on a body of water is presented. Starting from a field velocity distribution behind the vessel, the fluid mechanical behavior of the disturbance is developed as it interacts with the vertical thermal profile of the water and produces a temperature pattern on the sea surface. As a result, the signal sensed by a remote detection system is obtained. A detailed sensitivity study of the computational results is conducted by separately varying each of the main code variables. This method yields program input values that best fit the measurements. Using a bell-shaped velocity distribution as an input, the contrast radiant intensity of a natural macrowake generated by a catamaran is computed to be -225 kW/sr. This compares with a measured value of -298 kW/sr, from which it differs by about 24%. The variation of the spatial width of the macrowake on the water surface is computed and found to fit the measurements. An infrared gap is sometimes observed at the onset of the macrowake. Actual imagery of a catamaran and a monohull ship shows definite evidence of gaps. Their extent can be predicted by the theory to better than 5% accuracy.

The status of high-power (50 to 1000 mW) diode lasers for space communications is reviewed, with emphasis on monolithic devices. The current performance and future potential of single-element devices, few-element (two to three) phase-locked arrays, and many-element (> 10) phase-locked arrays are presented and discussed. A brief discussion of hybrid devices for achieving watt-range powers is presented as well. Subject terms: high-power diode lasers; phase-locked arrays; nonabsorbing-mirror diode lasers; quantum-well lasers; diffraction-limited beam; master-oscillator power amplifier; injection-locked diode lasers; direct detection laser-transceiver; coherent detection.

One of the primary duties Martha Stockton has as Associate Editor of Optical Engineering is to correspond with individuals I have identified as prospective referees for manuscripts we hace received. She initially inquires about their availability and willingness to serve as referees and then, depending on their response, follows up in some fashion.

This book is a collection of 16 chapters on various topics in optical signal processing, each written by one or more leading experts currently active in that field.