This special issue of Optical Engineering is concerned with visual communications and image processing. The increase in communication of visual information over the past several decades has resulted in many new image processing and visual communication systems being put into service. The growth of this field has been rapid in both commercial and military applications. The objective of this special issue is to intermix advent technology in visual communications and image processing with ideas generated from industry, universities, and users through both invited and contributed papers. The 15 papers of this issue are organized into four different categories: image compression and transmission, image enhancement, image analysis and pattern recognition, and image processing in medical applications.
A critical survey of and a current status report on transform coding technology is complemented by a review of additional potential improvements. The discussion provides an overview of transform coding technology and assesses the relative importance of the various algorithm classes. Topics discussed include adaptive procedures in both the spatial and transform domains, block size variations, and buffering. The importance of preprocessing is reviewed. The conclusions support the view that additional advances in the field are possible and are likely to occur.
Image compression using memoryless vector quantization (VQ), in which small blocks (vectors) of pixels are independently encoded, has been demonstrated to be an effective technique for achieving bit rates above 0.6 bits per pixel (bpp). To maintain the same quality at lower rates, it is necessary to exploit spatial redundancy over a larger region of pixels than is possible with memoryless VQ. This can be achieved by incorporating memory of previously encoded blocks into the encoding of each successive input block. Finite-state vector quantization (FSVQ) employs a finite number of states, which summarize key information about previously encoded vectors, to select one of a family of codebooks to encode each input vector. In this paper, we review the basic ideas of VQ and extend the finite-state concept to image compression. We introduce a novel for-mulation of the state and state-transition rule that uses a perceptually based edge classifier. We also examine the use of interpolation in conjunction with VQ with finite memory. Coding results are presented for monochrome images in the bit-rate range of 0.24 to 0.32 bpp. The results achieved with finite memory are comparable to those of memoryless VQ at 0.6 bpp and show that there are significant gains to be obtained by enhancing the basic VQ approach with interblock memory.
Progressive image transmission allows an approximate image to be built up quickly and the details to be transmitted progressively through several passes over the image. This technique appears very useful for picture communication over slow channels. This paper presents a survey of progressive image transmission techniques including spatial domain, transform domain, and pyramid-structured approaches. Image coding techniques that are not currently used for progressive transmission, but that have this capability or can be modified to provide progressive transmission, are also identified. The main features, computational complexity, and typical performance are discussed for each technique. Among various approaches, there are always trade-offs between system complexity and performance. Which approach is most appropriate depends upon the application.
This paper describes motion video compression transmission for teleconferencing at a subprimary rate, i.e., at 384 kbits/s, including audio signal through the integrated services digital network (ISDN) HO channel. A subprimary rate video coder/decoder (codec), NETEC-XV, is available commercially that can operate at any bit rate (in multiples of 64 kbits/s) from 384 to 2048 kbits/s. In this paper, new algorithms are described that have been very useful in lowering the bit rate to 384 kbits/s. These algorithms are (1) separation of moving and still parts, followed by encoding of the two parts using different sets of parameters, and (2) scene change detection and its application to encoding parameter control. According to a brief subjective evaluation, the codec provides good picture quality even at a transmission bit rate of 384 kbits/s.
Theoretical analysis of integrated local area network model of MAGNET, an integrated network testbed developed at Columbia University, shows that the bandwidth freed up during video and voice calls during periods of little movement in the images and periods of silence in the speech signals could be utilized efficiently for graphics and data transmission. Based on these investigations, an architecture supporting adaptive protocols that are dynamicaly controlled by the requirements of a fluctuating load and changing user environment has been advanced. To further analyze the behavior of the network, a real-time packetized video system has been implemented. This system is embedded in the real-time multimedia workstation EDDY, which integrates video, voice, and data traffic flows. Protocols supporting variable-bandwidth, fixed-quality packetized video transport are described in detail.
Perspectives of a few challenging problems in the digital processing of low resolution halftone images are presented. The topics covered include screen selection, tonal reproduction control, and the scaling and resolution conversion of halftone images. Particular emphasis is placed on a new method of digital scaling and resolution conversion of electronically stored low resolution halftone images. Several simulation examples are given.
Much of the work done in digital image processing has been limited in application to black-and-white images, this being especially true of enhancement and restoration. The extension to color image processing is not trivial; a suitable color space must be selected for a given application, and then a good processing strategy must be devised. In fact, we doubt that any of the available color spaces will meet the needs of all types of image processing. Many color image processing strategies require that only a luminance component be actually processed. In image restoration, for example, good results are achievable by processing only the Y component of the popular NTSC transformation from RGB to YIQ components. In this paper we show that color saturation, as well as luminance, can play an important role in achieving good image enhancement. The technique proposed is simple to implement and is based on the observation that the saturation component often contains high frequency components that are not present in the luminance component. Contrast and sharpness enhancement techniques are discussed; the computer processing algorithms are restricted to those that preserve the natural appearance of the scene. We also discuss limitations to luminance and saturation processing caused by poor quantization of the RGB tristimulus images.
We introduce a simple algorithm to reduce artifacts that often appear in image restoration techniques such as Wiener filtering. The algorithm starts with the inverse filter solution and iteratively calculates the correction term. At each iteration we use an entropy gradient and an analytically calculated step size. The algorithm uses two Fourier transforms per iteration. We show both 1 -D and 2-D examples to illustrate the algorithm.
This paper reviews some recent advances in the theory and applications of morphological image analysis. Regarding applications, we show how the morphological filters can be used to provide simple and systematic algorithms for image processing and analysis tasks as diverse as nonlinear image filtering, noise suppression, edge detection, region filling, skeletonization, coding, shape representation, smoothing, and recognition. Regarding theory, we summarize the representation of a large class of translation-invariant nonlinear filters (including morphological, median, order-statistic, and shape recognition filters) as a minimal combination of morphological erosions or dilations; these results provide new realizations of these filters and lead to a unified image algebra.
A model-constrained rule-based object recognition system for line drawings is considered. The system is composed of three subsystems: preprocessing (segmentation), feature extraction, and pattern matching. The subsystems and the processes within each system are arranged in a hierarchical structure. The actual implementation is centered around the ideas of view-angle-independent object recognition and robustness with incomplete or noisy line drawings of objects. The line drawing of a scene containing simple objects is first segmented into independent objects using a set of heuristics stored in the knowledge base. The segmented object is then mapped into a newly devised structural-symbolic representation that is independent of small view-angle changes. Large view-angle independence is achieved by multiple models for the same object. Incomplete line drawings and noisy line drawings are detected and modified by the preprocessing module before being presented to the pattern matcher. The matching process is carried out by production rule inferencing. The system implemented has the characteristics of being flexible and easy to modify. These advantages are rooted in the utilization of the rule-based approach. A test of the system has been carried out, with a satisfactory result.
An automatic recognition method to extract individual map components from large-scale maps in a digital format is proposed. This method recognizes the components on the basis of their geometric features and of drawing regulations of national large-scale maps of Japan. The geometric features are obtained by binary image processing techniques such as labeling, border tracing, line tracking, and thinning. At present, the method is capable of extracting buildings, railways, and contour lines. Experimental results have shown high extraction rates for the method.
This paper examines the classification potential of three techniques based on spiral sampling of gray-scale, noisy images. Image pixels are rearranged into a one-dimensional sequence by selecting samples in a spiral manner starting from the edge of the image and proceeding toward the center. The properties of this sample sequence are examined by Fourier transform and correlation techniques, using images from 26 groups with varying contrast, orientation, and size. The classification ability of features extracted from spiral sequences and their accuracy are investigated.
TOPICS: Visualization, Visual process modeling, Visual system, Algorithm development, Image enhancement, Signal to noise ratio, Computer simulations, Image processing, Fluctuations and noise, Point spread functions
The human visual system is capable of detecting and following the course of striated periodic patterns, even under adverse conditions of poor contrast and low signal-to-noise ratio. Sections of a striated pattern of subthreshold contrast may be detected easily if other parts of the same pattern have suprathreshold contrast. To simulate these capabilities of the visual system, an image processing algorithm was developed using basic "cells" that are well localized in both the space and spatial frequency domains. These band-limiting, orientation-sensitive "fan filters" are similar in their point spread functions to the two-dimensional Gabor functions commonly used to describe responses of visual cortical cells. These filters are used both to detect the orientation of the striated pattern in a small window and to enhance the image in that orientation. The search for local orientation is limited to a small range based on orientations found in neighboring, overlapping windows. The orientation of the maximally responding cell is used for the enhancement. Results of applying the adaptive directional enhancement to nerve fiber layer photographs, finger-prints, and seismic data are presented.
Two-dimensional gel electrophoresis is a powerful tool for determining the protein content of biological samples. At present, however, the great quantity of data to be extracted and examined does not allow its application on a wide scale; this result will be attained through the automation of the entire process. In this paper we analyze some basic problems involved in the automated analysis of electrophoresis images, propose some new solutions, and discuss the results obtained in experimental situations. In particular, spatial resolution is discussed, utilizing the results of spot-detection and segmentation procedures applied to images of three different resolutions. A spot-detection method based on the zero crossing of the first derivative of spot density is proposed. Spot segmentation is performed by testing the signs of second-order incremental ratios of spot density. Overlapped spots are separated by iteratively estimating the isolated spot densities, under the assumptions of density additivity and spot symmetry. Pattern recognition for feature extraction and cluster analysis for detection of spot gel constellations are proposed. Finally, an optical-flow calculation technique is modified for application to gel matching. The results obtained by means of such methodologies are promising and could be utilized by a knowledge-based system for automatic interpretation of electrophoresis images; this is the objective we propose for future research in this field.
Previous work has demonstrated the potential for adaptive filtration in processing digital chest images. The technique uses the histogram of the image to determine the pixels (and regions) in which edge enhancement is applied. This paper extends that work by investigating the choice of parameters used in selectively enhancing the mediastinum. The image is separated into its low and high frequency components by convolution with a square kernel. The effect of kernel size was studied with a choice of 17 x 17 mm, which was found to be sufficient to include the frequencies of interest. A serious deficiency in previous implementations of this technique is the existence of ringing artifacts at the juncture of the lung and mediastinum. These result in part from the use of a step function to specify the low frequency image intensity above which high frequencies are amplified. By replacing this step with a smoother (cosine) function, the artifact can be removed. Finally, the amplification constant was examined in light of its effect on both structure and noise in the image.
A new type of multilayer x-ray reflection coating has been de-signed and deposited: a so-called distributed Fabry-Perot etalon, consist-ing of a normal periodic multilayer in which extra periods of only spacer material are distributed. As an example, we deposited a ReW/C multilayer (d = 22 A, 90 periods) in which 90 extra C periods (d = 22 A) were ran-domly distributed. At X = 1.54 A, the resolution X/AX was limited by sub-strate flatness and was measured to be 133 (compared to about 90 for a , regular 90 period stack). The peak reflectivity is 21%. Crucial for deposition of this structure was a reduced substrate temperature (T = -150°C). The effective roughness of the interfaces is a constant 5 A throughout the entire stack.
When the Society of Photo-Optical Instrumentation Engineers (SPIE) began calling itself PIE hie International Society for Optical Engineering in 19811, an important motivating factor was the desire of the Society's Governors to reflect a rise in the membership from countries other than the United States and to foster an increase in cooperative activities with the optical engineering communities within those countries. At a time when the Society's technical focus had drifted away from photographic instrumentation and toward optical engineering, its membership had become less exclusively comprised of individuals from the U.S. and was beginning to develop a much more international flavor.
The soft x-ray region covers wavelengths
longer than those of medical x rays, where
vacuum is required for propagation (A ^ 5 A).
Absorption of all materials increases with
increasing wavelength in this region, and the
absorption length of, the most transparent
specimens is around 1 pm for A « 50 A.
Conventional lens and mirror systems cannot
be built for soft x rays (due to absorption
in lenses and low reflectivity of normal incidence
mirrors), and only three types of imaging
systems remain possible for soft x rays:
grazing incidence reflectors, multilayer mirrors,
and diffractive elements such as zone
plates. Soft x-ray optics has advanced dramatically
during the last decade.