Showing papers on "Image processing published in 1974"

Computer Processing of Line-Drawing Images

Abstract: This paper describes various forms of line drawing representation, compares different schemes of quantization, and reviews the manner in which a line drawing can be extracted from a tracing or a photographic image. The subjective aspects of a line drawing are examined. Different encoding schemes are compared, with emphasis on the so-called chain code which is convenient for highly irregular line drawings. The properties of chain-coded line drawings are derived, and algorithms are developed for analyzing line drawings to determine various geometric features. Procedures are described for rotating, expanding, and smoothing line structures, and for establishing the degree of similarity between two contours by a correlation technique. Three applications are described in detail: automatic assembly of jigsaw puzzles, map matching, and optimum two-dimensional template layout

Real-time correction of atmospherically degraded telescope images through image sharpening

Abstract: We present a new technique for the correction of atmospheric distortion in telescope images. Most of this distortion arises from a random phase variation of the incoming light across the telescope aperture. This variation limits the resolving power of even large telescopes to about one arc second. If the sharpness of the images is defined in a suitable way, this sharpness is maximized only when the phase distortion of the incoming light is zero. We present computer simulations of a simple feedback system in which active optical elements, set to maximize the sharpness, correct most of the atmospheric distortion. Photon statistics set the limiting magnitude of the object for which a practical feedback system can work. Details in a sixth magnitude object smaller than 0.1 sec of arc should be resolvable. The system can be conveniently employed within existing telescopes.

Geometric modeling for computer vision.

Abstract: : The main contribution of this thesis is the development of a three dimensional geometric modeling system for application to computer vision. In computer vision geometric models provide a goal for descriptive image analysis, an origin for verification image synthesis, and a context for spatial problem solving. Some of the design ideas presented have been implemented in two programs named GEOMED and CRE; the programs are demonstrated in situations involving camera motion relative to a static world.

Correlation Techniques of Image Registration

Abstract: An extension to the basic concept of correlation detection as a means of image registration is developed. The technique involves linear spatial preprocessing of the inages to be registered prior to the application of a correlation measure. This preprocessing operation utilizes the spatial correlation within each image and greatly improves the detectability of image misregistration. An analysis of the computational aspects of the algorithm is given. Also, results of a computer simulation to evaluate the technique are given.

A parametric model for human faces.

Abstract: : This report presents a computer model for the representation of human faces. This three-dimensional, parametric model produces shaded facial images. The face, constructed of polygonal surfaces, is manipulated through the use of parameters which control interpolation, translation, rotation and scaling of the various facial features. With this model, very little input information is needed to specify and generate a specific face with a specific expression. The model has been successfully used to produce a large variety of facial images and several animated sequences. The animated sequence illustrate the power of the model to change facial expression and conformation. Experience with the model indicates that fewer than 10 parameters must be manipulated to produce reasonable speech synchronized facial animation. (Author)

An analysis technique for biological shape. I

Abstract: An analysis technique for biological shape based on the concept of bending energy has been developed. The technique permits the calculation of the amount of energy (work) that would have to be expended to form typical biological shapes out of a linear, thin-shelled medium. In addition the development of this analysis procedure provides several new results concerning image processing including a sampling theorem for simply connected closed contours and a fast algorithm for calculation of the bending energy.

Computer matching of areas in stereo images.

Abstract: : This dissertation describes techniques for efficiently matching corresponding areas of a stereo pair of images. Measures of match which are suitable for this purpose are discussed, as are methods for pruning the search for a match. The mathematics necessary to convert a set of matchings into a workable camera model are given, along with calculations which use this model and a pair of image points to locate the corresponding scene point. Methods are included to detect some types of unmatchable target areas in the original data and for detecting when a supposed match is invalid. Region growing techniques are di9 for extend matching areas into regions of constant parallax and for delimiting uniform regions in an image. Also, two algorithms are presented to show some of the ways in which these techniques can be combined to perform useful tasks in the processing of stereo images.

Digital and optical reconstruction of images from suboptical diffraction patterns.

Abstract: Digital and optical reconstruction techniques are applied to synthetic holograms that are recorded at suboptical frequencies. The first section of the paper considers digital reconstruction that entails the application of the inverse diffraction transform to the diffraction pattern of an object illuminated with a suboptical source. Different cross sections of the object are displayed on a CRT in a sequential fashion. Image enhancement techniques are also employed in the process of digital reconstruction. The second section outlines a method for partially alleviating the longitudinal distortion that is inherent in optical reconstruction from synthetic holograms because of the difference between the recording and reconstructing wavelengths. The paper considers both the diffuse and nondiffuse illumination schemes and discusses the relative advantages and disadvantages of digital and optical reconstruction schemes.

Image Restoration, Modelling, and Reduction of Dimensionality

Abstract: Recursive restoration of two-dimensional noisy images gives dimensionality problems leading to large storage and computation time requirements on a digital computer This paper shows a two-dimensional second-order Markov process representation can be used for fast recursive restoration of images with small storage requirements Advantages of this method over existing techniques are illustrated by means of examples

Image Processing In Signal-Dependent Noise

Abstract: Image detection noise is a fundamental limitation in picture processing, whether analog or digital. This noise is characteristically signal-dependent and this signal-dependence introduces significant problems in the design of appropriate noise-suppression techniques. This paper outlines some recent results obtained by the authors in the optimum suppression of two types of signal-dependent image noise: film-grain noise and photoelectron shot noise. The work in grain noise suppression involves deriving the minimum-mean-square error Wiener filter for a new form of signal-dependent noise model suggested in earlier work by T. S. Huang. Implementation of these filters by either coherent optical or digital processing techniques is possible. Digital computer simulations of grain noise suppression using two particular cases of this additive, "signal-modulated" noise model were performed. They demonstrate the potential advantages of noise suppression filters which make use of a priori knowledge of the signal-dependent nature of the grain noise. The results of work on linear, unbiased restoration of images recorded in the presence of photoelectron noise are summarized. Additional work in both of these areas is suggested, with a particular need existing for correlating the properties of various models proposed for grain noise with experimental data obtained on emulsions using scanning microdensitome ters.

Digital image restoration: A survey

Abstract: The state of the art in large-scale digital computers has recently opened the way for high resolution image processing by digital techniques. With the increasing availability of digital image input/output devices it is becoming quite feasible for the average computing facility to embark upon high-quality image restoration and enhancement. The motivation for such processes becomes self evident when one realizes the tremendous emphasis man puts on his visual senses for survival. Considering the relative success achieved in one-dimensional (usually time) signal processing, it is to be expected that far greater strides could be made in the visual two-dimensional realm of signal processing.

Image Enhancement Techniques for Cockpit Displays

Abstract: : The major objective of this research program was to develop and evaluate three techniques for the real-time enhancement of sensor displays in the airborne cockpit environment. Two enhancement techniques developed and optimized for contrast enhancement are Local Area Brightness and Gain Control (LABGC) and Local Area Histogram Equalization (LAHE). Haar Transform Crispening (HTC) is another technique that was developed and optimized during this program. HTC may be used to sharpen edges in an image which are blurred due to misfocused optics or MTF roll-off. An evaluation of the enhancement techniques includes subjective evaluations for specific operator tasks and a comparison of mechanization complexities. Photographs of original and enhanced images are included.

The representation of two-dimensional sequences as one-dimensional sequences

Abstract: A number of signal processing techniques which have been developed for processing one-dimensional sequences do not generalize to the processing of two-dimensional signals, largely due to the absence of a two-dimensional factorization theorem. In an attempt to circumvent this problem, a specific representation of two-dimensional sequences as one-dimensional sequences is presented in this paper. Using this mapping several two-dimensional problems can be viewed as one-dimensional problems and approached using one-dimensional techniques. This representation is valid both for signals of finite extent and for the more general class of signals with rational Z-transforms. In this paper we consider applications of these techniques for high speed convolution, processing of drum scans, and two-dimensional finite impulse response (FIR) filter design.

Imaging Characteristics of the Itek PROM.

Abstract: This paper discusses advances in the fabrication, operation, and performance of the Itek PROM (Pockels readout optical modulator). Devices of high optical quality have been made and evaluated. Measurements of MTF and sensitivity under various conditions are presented. Stored images can be enhanced in contrast by a change in device voltage and the same method used to suppress the zero order in Fraunhofer diffraction patterns of such images. The results of this technique are demonstrated.

Quantitative Effects of Stationary Linear Image Processing on Noise and Resolution of Structure in Radionuclide Images

Abstract: Processing radionuclide image data in an attempt to improve its diagnostic usefulness inevitably changes both the noise and the resolution of structures in the image. Although at present little conclusive evidence exists to indicate whether or not image processing is beneficial, an understanding of the quantitative effects of various processing techniques on parameters of resolution, noise magnitude, and noise texture seems to provide a useful beginning towards solution of the more difficult problem of understanding the effects of image processing on diagnostic image quality. Quantitative effects of stationary linear image processing procedures, which include all convolutions and filtering operations, can be predicted rather simply. In the present work, expressions for predicting these quantitative effects are developed and are applied to computer-synthesized image data to provide illustrative examples of the effects of typical noise-smoothing and resolutionenhancing operations. (auth)

Variable intensity display device

Abstract: A large display system capable of displaying a video image receives video signals, quantizes those signals to produce a digital code capable of representing variations in the light content of the image and processes the digital code so as to control individual display devices on a large matrix of such devices to have different levels of visibility to thereby reproduce the video image for viewing by a large audience. A data processor is optionally utilized to store the digital representation of the video image in memory so that on line or off line presentations can be made.

Correction of spatial non-uniformities in sonar, radar, and holographic acoustic imaging systems

Abstract: An improved method and apparatus for processing electronic signals for usen establishing an improved image display includes storing the electronic and acoustic response of all the detectors in response to impingement thereon by a plane wave-front as stored in a memory. In the operational environment, these stored signals are combined with the transducer response and the signals are combined in such a fashion as to compensate for spatial nonuniformities.

The Use Of A General Description Of The Radiological Transmission Image For Categorizing Image Enhancement Procedures

Abstract: The mathematical function describing the transmission of x-rays through an object depends on many geometrical and physical variables. A Taylor series expansion of the transmission function about an arbitrary point in the multivariable space displays the variety of terms which may be emphasized or isolated in order to accentuate certain information concerning the object under study. Conventional radiographic or fluoroscopic images are represented by the zero order term. This term can be used to derive higher order terms which depend explicitly on the spatial coordinates x and y, but contains only averaged or incomplete information about variables such as z (depth), E (energy), or t (time). One class of image enhancement procedures involves accentuation of whatever information is present on conventional radiographs by selective filtration of spatial frequency components or isolation of the x and y derivative terms. A second class of images with significantly different information content can be obtained by a priori use of knowledge about other variables in order to isolate other derivative terms. These images can be formed from linear combinations of zero order images associated with different values of the variable of interest. Examples of this class of images include tomography, time dependent subtraction, and absorption edge imaging. Because evaluations of "enhanced" images containing one subset of the possible image terms cannot be assumed to pertain to images containing other terms, it is suggested that the term "image enhancement" is too general and must be used with care.

Method and apparatus for combining video images with proper occlusion

Abstract: An image generated by a television camera is combined with an image generated from digitally stored information to form a composite scene wherein objects appearing in either image properly occlude one another with respect to the eyepoint from which the scene is viewed. This is accomplished by processing the camera video signal to provide pulses for each raster line which indicate the starting and stopping points of an object viewed by the camera. At the same time, the camera video signal is provided on a delay line to a video switch which also receives the digitally generated video signal. The switch is controlled by logic circuitry in accordance with the established occlusion priority to pass to the video display device the proper signal from either the camera or the image generator.

Signal correction system

Abstract: In a system producing stored digital words representing an optical image, a digital error correction system automatically corrects the video signal for shading. A television camera converts the optical image of a laboratory microscope slide into an analog electrical signal representing the optical image along a raster of scan lines. An analog-to-digital converter produces digital words representing the magnitude of the video signal at periodic times along each scan line. These words are stored to represent the optical characteristics of the image. In order to correct the analog video signal for shading error, digital words representing the pattern of the shading are stored in a digital memory. This memory is loaded during intervals in which the analog signal represents only the shading. During this time, the video signal is converted to digital words and stored in the digital memory. Thereafter, during normal operation, these digital words are converted to an analog error signal which is subtracted from the video signal.

Complex filters, convolvers, and multipliers

Abstract: A complex multiplier, using only real multipliers, having two pairs of signal inputs, at one pair of which appears the real and the imaginary parts, A and B, of an arbitrary signal A + jB, which had been decomposed into these components prior to appearing at the pair of inputs, at the other pair of signal inputs appearing the real and imaginary parts C and D, of a similarly decomposed, arbitrary, signal C + jD. The complex multiplier comprises four signal summers, two means for inverting a signal, and three signal multipliers. The magnitude of the output signal of the third summer is equal to the magnitude of the imaginary part of the product of the complex signals, A + jB and C + jD, while the magnitude of the output signal of the fourth summer is equal to the magnitude of the real part of the same multiplied complex signals.

The Prom Device In Optical Processing Systems

Abstract: The operation of an electro-optic image modulator device for use in real-time optical processing systems is described. The PROM (Pockel's Read-Out Optical Modulator) is used to provide temporary storage of image or digital data which can then be read out by a laser for coherent optical processing. The stored image may undergo certain operations, such as level slicing, edge enhancement, or contrast inversion by utilizing the electronic characteristics of the device. This feature is useful for incoherent image processing, and also allows zero order suppression in coherent processing applications. The rapid recyclability, high optical quality and high resolution demonstrated by the device are required for practical utilization of the device in the applications described.

Reduction of the number of samples in computer holograms for image processing.

Abstract: Institut d'Optique, Université de Paris XI, Centre d'Orsay, 91405 Orsay, France. Received 14 December 1973. Computer generated holograms improved significant­ ly the possibilities of optical data processing by spatial fil­ tering. They allow an easy synthesis of several inter­ esting filter functions for the purpose of deconvolution and other image enhancement methods. Nevertheless, a practical limitation appears when the processed image has to contain a large number of degrees of freedom, i.e., when a large space-bandwidth product is involved. Then, in order to avoid overlapping of the vari­ ous orders in the image plane, the number of sampling cells of the computer generated Fourier hologram used for filtering must be at least equal to the number of degrees of freedom of the image to be processed. As an example, let us consider the case of a 24 × 36 mm image recorded with an f/18 lens at a wavelength of 0.5 μm: it contains a number N0 = 10 degrees of free­ dom. The processing hologram has to contain this num­ ber of sampling cells and needs a large computing and plotting time. Moreover, the plot is not easy to perform because, even with a cell size as small as 1 mm, the plot area is 10 m or more. In addition, the requirements upon the optical reduction system are very severe. The trouble is that these conditions hold even when the im­ pulse response of the hologram contains only a small number of degrees of freedom, say 10. This means that a large oversampling is necessary, not to perform the desired processing but only to avoid image overlapping. A typical example is pattern recognition of a small detail in a large image. Another example is relat­ ed to images degraded by some small blurring function. Then the deblurring function is in general of the same physical size as the blurring function, i.e., it contains many less degrees of freedom than the image. Our aim is to elude this difficulty and to compute and plot a binary hologram with only the minimum number of resolution cells that allow reconstruction of a given im­ pulse response with a given bandwidth, irrespective of the number of image resolution cells. To this end, we pro­ pose a two-step method. 1. A binary computer hologram (C.H.) with a small num­ ber of cells is realized. The space-bandwidth of this hologram is just large enough to prevent overlap of ad­ jacent impulse response orders. 2. A natural ordinary hologram (N.H.) is recorded using one order of (C.H.) as the object. By adjusting proper-

Scanning line filter

Abstract: An image processing device for enhancing and digitizing images made up primarily of continuous parallel lines such as contour maps and fingerprints. An image is focused on a photodiode detector array and a digital logic signal is produced which is the binary representation of the portion of the image centered on the array.

Discrete cosine transform signal processor

Abstract: A processor for performing a discrete cosine transform of an input signal, suitable for real-time television image processing, specifically for obtaining an acceptable picture when the number of bits of information available for describing the picture and/or the channel bandwidth are severely limited, comprising: two complex read-only memories, an input and output read-only memory, each containing a predetermined number of data points arranged in a predetermined manner; two complex multipliers, an input and an output multiplier, each having an input from one of the read-only memories, an input which is connectable to the external signal of N data values which is to be transformed discretely and cosinusoidally; a complex transversal filter, having 2N-1 taps, the input to the filter being the output of the input multiplier; the output of the transform processor comprising the output of the output multiplier.

Digital Image Deblurring by Nonlinear Homomorphic Filtering

Abstract: : This report is concerned with the digital estimation of the frequency response of a two-dimensional linear system through which images have been passed and blurred. Almost no a priori knowledge concerning the system is required, and only one blurred image is necessary for a successful estimation. For those blurs that have phase reversals, such as motion blurs and out-of-focus blurs, a simple method of calculating the correct phase has been developed. A nonlinear homomorphic restoration system is developed and demonstrated on various types of blurred images. An image may be restored by filtering either the intensity version or the density version of the image. The latter not only insures a positive result but also permits simultaneous deblurring and contrast enhancement. The restoration procedure consumes only a modest amount of computation time.

The shadow transform: An approach to cross sectional imaging

Abstract: : The goal of this work is to define and investigate the properties of the linear operator that relates the density to its set of projections--the shadow. Several approaches to the definition of the shadow transform are presented in Chapter I. Chapter II provides the basis for the spectral analysis by deriving the eigenfunctions of the shadow transform. Chapter III deals with the derivation of reconstruction schemes (realizations of the inverse shadow transform). In Chapter IV the author develops the theory for the case where the measurement of the projections is limited by the geometry.

Digital image processing at EG&G

Abstract: EG&G, Inc., a leader in the field of photographie data acquisition and analysis for over 25 years, has for the past five years been placing increased emphasis on digital image processing in support of the Field Testing Division of the Los Alamos Scientific Laboratory under AEC Contract No. AT(29–1)1183. During that time, capabilities that originally involved photogrammetric and radiometrie analysis have been augmented to include two-dimensional Fourier frequency analysis and spatial filtering of images. This article describes some of the image processing equipment at EG&G, along with the results obtained using this equipment, and discusses digital holography in greater detail.