A raster based variable focal length (varifocal) mirror display system, which is integrated into a high-speed 200 million instructions per second (MIPS) image processor (IP), is described. The display system is comprised of a loudspeaker-membrane-mirror subassembly, a suitable cathode ray tube (CRT) display monitor, and a display controller. The display controller is a single circuit board installed directly into the chassis of the IP. The stack of tomographic images, to be displayed as a three-dimensional raster, is stored in the 32 megabyte IP memory. Both the display controller and the computing engine of the IP have direct access to the image arrays over very high speed data pathways. The varifocal mirror display system uses a vibrating mirror in the form of an aluminized membrane stretched over a loudspeaker, coupled with a CRT suspended face down over the mirror. The mirror is made to vibrate back and forth, as a spherical cap, by exciting the loudspeaker with a 30 Hertz sine wave. Stacks of 2-D tomographic images are displayed, one image at a time, on the CRT in synchrony with the mirror motion. Because of the changing focal length of the mirror and the integrating nature of the human eye-brain combination, the time sequence of 2-D images, displayed on the CRT face, appears as a 3-D image in the mirror. The system simplifies procedures such as: reviewing large amounts of 3-D image information, exploring volume images in three dimensions, and gaining an appreciation or understanding of three-dimensional shapes and spatial relationships.
The SpaceGraph technology has been packaged as a table-top display station that is a peripheral device for an IBM PC/AT or compatible. It produces 3-D graphics by oscillating the virtual image of a CRT through space while synchronously writing on the screen. In the resulting display volume, one sees a self-luminous, high-contrast, sharp, model-like figure composed of points, lines, and alphanumerics. Optical means include a varifocal plate mirror and a specially designed, large-screen, directed-beam monitor. Electronic means include a controller on a card for an IBM PC/AT or compatible. Software running in the PC allows one to describe a 3-d picture in high-level terms. The controller card can display two pictures at once, and each is double-buffered. A run-time-writable brightness-lookup table takes advantage of 8-bit brightness tags on all 32k displayable points. The software interface is command-driven and includes general and command-specific help. It is capable of getting commands either from files, from the keyboard, or from a combination. Commands take intuitive forms, such as line (3,4,5) (1,2,6) or text (2,4,3) "abc" and locations in space may be given names for convenience of future reference. Additional features are autoscaling, a cursor, and saved display lists. Existing application software can most easily interface to SpaceGraph by writing out files of such commands. An example is the supplied interface to BBN's proprietary software product, RS/1.
Parallactiscopes (parallactic oscilloscopes) produce holoform (hologram-like) images in real time by reconstructing ray directions with a horizontally-scanned vertical slit. The slit causes each narrow vertical zone on the CRT screen to be keyed to a unique viewing direction, thus presenting a different total image to each of the many eyes of a group of observers. As a direct result of the process, stereo and movement parallax are produced automatically and without any observer constraints or observer-worn hardware being necessary. Solid models (opaque images) can be displayed as well as wire-frame (transparent) models. Viewing angles up to Ã‚Â±45 degrees off axis are readily accommodated. This paper describes the latest in a line of parallactiscopes designed, built, and demonstrated by the author. It incorporates a 19-inch electrostatically-deflected CRT.
This paper outlines the development of a new method of 3-D imaging based on spatial image decomposition by ray direction. Various implementations of actual systems and future possibilities will be discussed, including full color auto-stereoscopic video and synthetic holography of real and computer generated objects. Practical and theoretical limitations will be addressed, as well as bandwidth limitations and redundancy. The adaptation of standard methods of bandwidth reduction by adaptive image coding and processing will be considered. The integration of the new technology with computer graphics and holography will also be examined within the context of a "window" theory of spatial imaging.
A new type of large liquid crystal modulator, the ZScreen(R), has been developed for use in a stereoscopic display system and has been put into production by StereoGraphics Corporation. The fabrication of large area surface mode liquid crystal cells requiring extreme uniformity of thickness has been achieved in a two year development program. Two of these cells are placed in optical series, with their alignment axes orthogonal. Laminated to the liquid crystal cells is a linear polarizer whose absorption axis besects the alignment axes of the liquid crystal cells. When placed in intimate juxtaposition with a CRT display with the linear polarizer facing the CRT screen, and when the cells are driven electrically out of phase, the light output of the CRT is alternately left and right handed circularly polarized. When this so-called push-pull modulator is viewed through passive circularly pollarized spectacles worn by an observer, a shutter is formed having a high dynamic range. Each eye of the observer is sequentially occluded. By properly choosing the parameters of construction, such a liquid crystal modulator can have a broad cone of view, high dynamic range, and good transmission. This type of modulator is superior to a single liquid crystal cell modulator since it has fast and symmetrical rise- and decay times and identical dynamic range for both eyes. These features are of particular importance in a field-sequential stereoscopic system, because the rise and decay time of such a shutter should take place substantially within the vertical blanking interval, which is generally on the order of a millisecond. The 9-inch diagonal versions of the modulator have been in production since the last quarter of 1986 and now are being supplied in up to 19-inch diagonal sizes for computer graphics and video displays.
A chromostereoscopic display consists of a single flat image in which the depth information has been encoded into color. The image is viewed through a pair of chromostereoscopic spectacles, each lens of which consists of one or more superchromatic prisms. The superchromatic prisms are designed to maximize the chromatic dispersion so that different colored objects are given different amounts of horizontal parallax while minimizing the amount of angular deviation of the line of sight so that colors with median wavelengths appear at the plane of the screen. Our purpose in this work was to investigate the possibility of CRT-based chromostereoscopic display. We were particularly interested in how many depth levels we could distinguish and the effect of RGB color combinations. Four types of experiments were conducted: 1) depth effect for nonoverlapping objects of different colors in an image; 2) depth effect for overlapping objects of different colors; 3) the depth effect produced by gradually modifying the shading of a single object from red to blue; and 4) relative depth position of the base phosphor colors and color combinations.
Although new display technologies have reduced the cost and improved the quality of stereoscopic 3-D video displays, the question remains: when is the extra cost of a 3-D system justified by improved task performance? This paper will discuss: generic characteristics of tasks that are likely to benefit from 3-D techniques, common misconceptions about 3-D display limitations, problems of properly evaluating cost/benefit tradeoffs, and often-overlooked side-benefits of stereoscopic systems.
A comparison was made between viewing normal television and VISIDEPTM television which produces three-dimensional images by the method of alternating images. Two separate groups of fifteen university students reviewed fifty minute unrelieved exposure to television; one group watched standard television and the other watched VISIDEP. Both groups were surveyed regarding questions of eye strain, fatigue, headache, or other discomforts, as well as questions of apparent depth and image quality. One week later the participants were all shown the VISIDEP television and surveyed in the same manner as before. In addition, they were given a chance to make a direct side-by-side comparison and evaluate the images. Analysis of the viewer responses shows that in relation to viewer comfort, VISIDEP television is as acceptable to viewers as normal television, for it introduces no additional problems. However, the VISIDEP images were clearly superior in there ability to invoke an enhanced perception of depth.
This paper reports a human-factors evaluation of cursor-positioning devices for 3-dimensional (3-D) display workstations. Five input devices were compared in terms of cursor-positioning time and error with two 3-D display systems: a true 3-D stereoscopic display system and a conventional display system using linear perspective to convey depth information. For both 3-D display systems, the results indicated that vector-oriented input devices produced lower cursor-positioning errors and faster cursor-positioning time than either plane-oriented or free-space input devices. Several implementations of vector-oriented input devices are discussed.
A need for advanced design tools has motivated the design and development of three new concepts in device input for 3D displays: 1) a 3D sketching environment using 2D perspective input techniques, 2) B-spline approximation of free-hand curves in 3D using a 3D input device and stereoscopic 3D output and 3) an interactive cuberille sculpting environment supporting rapid presentation and modification of 3D objects.
The long term goal of the project described in this paper is to improve local tumor control through the use of computer-aided treatment design methods that can result in selection of better treatment plans compared with conventional planning methods. To this end, a CAD tool for the design of radiation treatment beams is described. Crucial to the effectiveness of this tool are high quality 3D display techniques. We have found that 2D and 3D display methods dramatically improve the comprehension of the complex spatial relationships between patient anatomy, radiation beams, and dose distributions. In order to take full advantage of these displays, an intuitive and highly interactive user interface was created. If the system is to be used by physicians unfamiliar with computer systems, it is essential that a user interface is incorporated that allows the user to navigate through each step of the design process in a manner similar to what they are used to. Compared with conventional systems, we believe our display and CAD tools will allow the radiotherapist to achieve more accurate beam targetting leading to a better radiation dose configuration to the tumor volume. This would result in a reduction of the dose to normal tissue.
The advances in stereoscopic image display techniques have shown an increased need for real-time interaction with the three-dimensional image. We have developed a prototype real-time stereoscopic cursor to investigate this interaction. The results have pointed out areas where hardware speeds are a limiting factor, as well as areas where various methodologies cause perceptual difficulties. This paper addresses the psychological and perceptual anomalies involved in stereo image techniques, cursor generation and motion, and the use of the device as a 3D drawing and depth measuring tool.
We have developed a "Stereoscopic Imaging System" (SIS) which acquires stereo pairs from video cameras (or other RS170 sources), aligns and enhances the images as needed, stores and recalls them from disk, displays them in "true" 3-D on a stereo monitor, and provides 3-D measurements of objects and features in the composite image. The system is based on the Tektronix Stereoscopic Graphics System (Liquid Crystal Shutter SPU, Tektronix, Inc.). This system uses a screen-sized liquid crystal to modulate the polarity of the light coming from the monitor at a rate of 120 times per second, and the left and right views are sent to the monitor at the same rate (each view is displayed at 60 Hz). The viewer wears lightweight, passive glasses with left and right circularly polarized lens. The display is controlled by an AT-compatible graphics card (the Stereoscopic Graphic Adapter or SGA) that stores each image with a resolution of 512x512x8, has room for four images (two pairs), and drives an RGB color monitor (64 KHz). (SIS uses color for overlays, but it does not yet handle color images.)
As real-time performance of 3D graphics machines is achieved the ability to interact more naturally with synthetic 3D scenes is needed. Humans manipulate objects in a spatial environment most easily with their hands and the same should be true for synthetic 3D scene manipulation. Thus, a system is needed for recognition of human hand position and configuration which does not encumber hand movement. This paper describes a system which computes a 3D reconstruction from multiple camera views of a human hand. .A polyhedron synthetic model of the human hand is then modified to most closely match the size and shape of the reconstructed human hand. Ulti-mately, the calibrated hand model will be used in a real-time model-based hand configuration recognition system. This system will include multiple cameras and back-lighting surrounding a convenient hand working space of perhaps one cubic meter. In order to manage the combinatorial explosion of all possible hand configurations, the hand model is divided into components (fingers, thumb and palm) and all possible configurations of the individual components are stored. Hand configuration recognition will be achieved by hierarchical partial matches of hand model components to the reconstructed user's hand. The search for the best matching hand model configuration will be limited by assuming a maximum possible range of hand component movements from frame to frame. Preliminary results from model-based hand configuration recognition experiments will be shown and methods for real-time implementation of the system will be discussed.
A technique for making holographic stereograms has been developed in which photographic film is replaced with a liquid crystal spatial light modulator to serve as an electronically controlled projection mask in the holographic recording process. The image source can be either a television camera and video cassette recorder that has recorded real scenes or a computer displaying three-dimensional computer graphics.
Several algorithms have been developed for Cartesian hyperspace graphics. These include 1) a hidden-line algorithm, 2) a hidden volume algorithm, 3) a shadow algorithm, 4) a shading algorithm, and 5) a raytracing algorithm. Algorithms are being developed for hextree representation. Related efforts include a study of perception cues for n dimensions. Objects in Cartesian hyperspace have been displayed using bimodal, dynamic stereo projection and holography.
The original data is produced through standard magnetic resonance imaging (MRI) procedures with a surface coil applied to the lower back of a normal human subject. The 3-D spine image data consists of twenty-six contiguous slices with 256 x 256 pixels per slice. Two methods for visualization of the 3-D spine are explored. One method utilizes a verifocal mirror system which creates a true 3-D virtual picture of the object. Another method uses a standard high resolution monitor to simultaneously show the three orthogonal sections which intersect at any user-selected point within the object volume. We discuss the application of these systems in assessment of low back pain.
A window manager for a volumetric workstation is under development. Based on experience with displaying transparent cubes of sampled data, the window manager treats windows as transparent PAGEs, and allows multiple PAGEs to be displayed simultaneously on a stereoscopic display. The users perceive this volume space and interacts with it with keyboard, voice input, mouse and 3 dimensional (3D) pointer. The overall design of the volumetric workstation and results from preliminary studies with transparent PAGEs are presented.
An automated segmentation algorithm for the isolation of pseudoinvariant features was developed as a part of this study.1 This algorithm utilizes rate-of-change information from the thresholding process previously associated with the pseudoinvariant feature normalization technique demonstrated by Volchok and Schott, 1986.2 The segmentation algorithm was combined with the normalization technique and applied to the six reflective bands of the Landsat Thematic Mapper (TM) for both urban and rural scenes. The technique was also applied to color infrared high resolution U2 imagery. The accuracy and precision of the normalization results were evaluated. The combined techniques consistently produced normalization results with errors of approximately one to two reflectance units for both the rural and urban TM imagery as well as the visible bands of the high resolution air photo imagery. These results compare favorably with previous findings utilizing the manual segmentation technique while simultaneously eliminating user-to-user inconsistencies. Also developed as a result of this study was a quantitative metric for comparison of different normalization techniques.
This is a pilot demonstration to evaluate the overall biological system related to the jaws, mouth, and face (i.e., stomatonagthatic system). These elements will be evaluated as to their biomechanical function, imaging of their anatomical parts, and examining the biological reactivity of the system. This will be presented in a display of multi-modalities and accessed by an expert system shell.
We are developing a system designed around an IBM PC-AT to perform automatic diagnosis of diseases from images of the retina. The system includes hardware for color image capture and display. We are developing software for performing image enhancement, image analysis, pattern recognition and artificial intelligence. The design goal of the system is to automatically segment a digitized photograph of the retina into its normal and abnormal structures, identifying these objects by various features such as color, size, shape, texture, orientation, etc., and ultimately to provide a list of possible diagnoses with varying degrees of probability. We will discuss algorithms used to identify markedly different objects and to distinguish between those objects which appear very similar to the trained eye. Implementation of these algorithms, which are typically applied to areas such as remote sensing, terrain mapping and robotics, has been very successful when applied to color images of the retina.
Remote sensing applications continue to place increasingly strict specifications on laser systems. High pulse.energy and short-duration pulses relate directly to range and spatial resolution capabilities Of laser-based atmospheric sensors. Frequency agility of the laser adds the dimension of differential absorption and differential scattering measurements for quantitative studies of atmospheric molecules and aerosols. High spectral purity of the laser transmitter pulse permits the use of heterodyne detection for improved signal-to-noise measurements and makes the studies of small Doppler-induced frequency shifts caused by relative motion between target and observer possible. Laser Science, Inc. has built a high-spectral purity, injection locked CO2 laser transmitter for remote sensing and target ranging applications. The purpose of this paper is to discuss the design details and the theory of operation of this laser transmitter.
Two information. extraction methods were applied to selected diagnostic medical images obtained from different imaging. modalities. Sample images consisted of magnetic resonance tomograms, transmission computed tomograms, and emission computed tomograms of patients with focal and generalized brain abnormalities. Color composite analysis applied to image pairs afforded a limited but potentially useful approach to enable rapid, objective. depiction of relative and absolute contribu-tions from dualmodality parameters. Cluster analysis applied to image pairs or triplets was computationally more demanding, but this approach afforded a method for objective classification of tissues that could be generalized. Technical problems arise from patient factors and machine factors. Patient factors include motion and positioning. Machine factors include resolution, orientation, format, registration and scaling.
Magnetic Resonance Imaging (MRI) is a relatively new diagnostic imaging modality that is rapidly finding broad clinical application. MRI differs from other diagnostic techniques in its capacity to obtain multiple qualitatively different images of the same anatomic region each emphasizing a different fundamental parameter of the tissue. This multiparametric nature of MRI provides the potential for greatly improved sensitivity and specificity in the detection of pathological conditions. However, the complexity of MRI can produce a potentially overwhelming volume of image data for the physician to analyze visually utilizing the traditional grey-scale. Additionally, "visual synthesis" of images from multiple data sets is only semi-quantitative at best and subject to errors introduced by observer bias. Data dimension reduction techniques are needed for analysis of these image sets of multi-parametric MRI data. It is hoped that improved diagnostic specificity of MRI will come from such a quantitative analysis of multiple MR images. Our initial experience with application of fuzzy clustering analysis to these MR images as a method of data dimension reduction suggests that such an approach can improve tissue specificity.