Development of a general purpose and practical 3-D display system appears feasible through the combination of fast data processing and transfer techniques with modern display methods. The technological problems and methods for effecting a merger of strategies and devices are discussed in this article.

The efficient representation and synthesis of 3-D object information using new synthetic discriminant functions (SDFs) is discussed. The use of SDFs in a correlator for shift-in-variant and distortion-invariant discrimination of 3-Dobjects is detailed and experimental data is provided. The new SDF3 described control the peak intensity and the structure and statistics of the correlation plane pattern.

In this paper we present some experiments in processing and display of true 3D-images (density volumes). The experiments have been carried out on the PICAP II image processing system.

The medical motivation for interactive manipulation of 3D data by a 2D display device is presented. Four specific topics are described: (i) interactive orientation specification, (ii) 3D83, a program currently used in clinical practice, (iii) directed contour based display and manipulation, and (iv) binary array based display and manipulation. Timings are reported on medical examples. Advantages and disadvantages of the various approaches are discussed.

A system for the three-dimensional (3-D) display and analysis of stacks of tomographic images is described. The device utilizes the principle of a variable focal (vari-focal) length optical element in the form of an aluminized membrane stretched over a loudspeaker to generate a virtual 3-D image which is a visible representation of a 3-D array of image elements (voxels). The system displays 500,000 voxels per mirror cycle in a 3-D raster which appears continuous and demonstrates no distracting artifacts. The display is bright enough so that portions of the image can be dimmed without compromising the number of shades of gray. For x-ray CT, a displayed volume image looks like a 3-D radiograph which appears to be in the space directly behind the mirror. The viewer sees new views by moving his/her head from side to side or up and down. The system facilitates a variety of operator interactive functions which allow the user to point at objects within the image, control the orientation and location of brightened oblique planes within the volume, numerically dissect away selected image regions, and control intensity window levels. Photographs of example volume images displayed on the system illustrate, to the degree possible in a flat picture, the nature of displayed images and the capabilities of the system. Preliminary application of the display device to the analysis of volume reconstructions obtained from the Dynamic Spatial Reconstructor indicates significant utility of the system in selecting oblique sections and gaining an appreciation of the shape and dimensions of complex organ systems.

In the practice of radiology, computer graphics systems have become an integral part of the use of computed tomography (CT), nuclear medicine (NM), magnetic resonance imaging (MRI), digital subtraction angiography (DSA) and ultrasound. Gray scale computerized display systems are used to display, manipulate, and record scans in all of these modalities. As the use of these imaging systems has spread, various applications involving digital image manipulation have also been widely accepted in the radiological community. We discuss one of the more esoteric of such applications, namely, the reconstruction of 3-D structures from plane section data, such as CT scans. Our technique is based on the acquisition of contour data from successive sections, the definition of the implicit surface defined by such contours, and the application of the appropriate computer graphics hardware and software to present reasonably pleasing pictures.

The combination of user-friendly, highly interactive software, 3D graphics, and the high-resolution detailed views of anatomy afforded by X-ray computer tomography and magnetic resonance imaging can provide surgeons with the ability to plan and practice complex surgeries. In addition to providing a realistic and manipulable 3D graphics display, this system can drive a milling machine in order to produce physical models of the anatomy or prosthetic devices and implants which have been designed using its interactive graphics editing facilities.

For the therapy of diseases of spinal deformity such as scoliosis, the data of 3-dimensional and correct spinal configuration are needed. Authors developed the system of spinal configuration analysis using bi-plane X-ray photogrammetry which is strong aid for this subject. The idea of correction angle of rotation of vertebra is introduced for this system. Calculated result under this idea has the clinical meaning because the correction angle is the angle which should be corrected on the treatment such as operation or wearing the equipment. Method of 30° oblique projection which gives the apparent X-ray image and eases the measurement of the anatomically characteristic points is presented. The anatomically characteristic bony points whose images should be measured on a- or b-film are of four points. These are centers of upper and lower end plates of each vertebra the center is calculated from two points which are most distant each other on the contour of vertebral end plate ), the lower end points of root of right and left pedicles. Some clinical applications and the effectiveness of this system are presented.

The Intelligent Task Automation Project involves the automatic location and assembly of a tray of parts in an uncontrolled. environment (i.e., no control of hunting). This paper provides a. brief description of the ACRONYM vision system and concentrates on extensions for location of the assembly components. The characteristics of the components are significantly different than the domains in which ACRONYM has been demonstrated -in particular the parts include holes and springs. Problems in representation, image feature prediction and image interpretation are discussed along with significance for future work in model-based vision

MRSD, which is the expanded concept of MRTD in passive thermal imager, was proposed to evaluate the system performance of an active coherent imaging system. An evaluation function of MRSD is achieved and a measurement system using computer generated 4-bar target is considered. Several results of calculated and measured MRSD are fitted well.

This paper describes the design of a three-dimensional display device capable of display-ing real-time dynamic images. The system combines the power of pre-processing Yith the strengths of a particular high-bandwidth display configuration to produce dynamic high-bandwidth volumetric images. Emphasis in this paper is given to design aspects required to provide the moving image capability to an existing static 3-D display system.

An implementation of a true 3-dimensional display-system using vibrating mirror is described. This implemen-tation is characterized by using a standard raster graphics system, and supports user-interaction for independent manipulation of multiple objects that is near real-time. The high-speed performance of the object-manipulations, which include translation, rotation, scaling, intensity-windowin g, spatial-clipping, intensity-highlighting, and blinking, is achieved through five means: exploiting the power of the user-programmable graphics processor, fully utilizing the raster-device characteristics, coordinating con-current execution of the host and graphics computers, localizing the image changes effected by interaction, and by relying upon successive refinement of the initially coarse object-display during periods of reduced interaction. Recent enhancements include the capability to display more points (circa 120,000), the integration of the functions into one system allowing multiple simultaneous manipulations, and the support for independent object motion. The system is currently being used to display CT and NMR data for medical imaging and electron microscope tomo-graphs for molecular modeling.

A simple approach to achieving visually effective stereoscopic presentations on CRT display systems using electronically controlled shutters has been developed and refined. This method employs lead lanthanum zirconate titanate (PLZT) ceramics with dichroic polarizers as electro-optic shutters which are driven synchronously with the refresh rate of a CRT. The principal advantage of the approach presented here is the ease with which strong three-dimensional visual effects can be produced using basic stereoscopic viewing equipment in conjunction with standard video components or conventional computer-driven CRT display hardware. Applications of the basic PLZT electro-optic shutter stereoscopic video imagery generation technique to various display systems as well as methods of generating stereoscopic video and computer-driven display system imagery are presented.

In the course of investigation, the author determined a set of criteria for establishing the successful transmission of stereoscopic images. These criteria are termed binocular symmetries. A point-for-point correlation is established between the left and right image fields. This correlation is defined as a symmetrical quality, which can be analyzed into a number of categories applicable to the assessment of the images.

Generally described in this paper are the opto-electronic techniques for the feedback control of automated equipment. Both spatially and temporally based architectures are used separately and in hybrid configurations. The present project concerns the six-dimensional topographic reading of plates and the weld seam between them.

Holography has long been considered to be an ideal method of visually presenting three-dimensional data which has been either experimentally acquired or mathematically calculated with a preconceived form in mind. Unfortunately, holograms have not had the image quality and accuracy to gain scientific acceptability outside the field of interferometric analysis. This paper will describe the use of an advanced composite holographic technique to give a clear, autostereoscopic image of a computer designed automotive part and a computer graphics concept of a molecule. The holograms were manufactured using Image Plane Integral (IPI) holography on an Argon laser printer designed and built by the author. The final holograms exhibited bright, achromatic, real images with high resolution and a minimum amount of controllable distortion. Additionally, because the IPI method uses a 35mm microfilmstrip for input, it could be universally applied as a hard copy format for presentation of almost any type of three-dimensional data base.

During the 1981 summer season within a 70,000 km2 area surrounding Miles City, Montana, the meteorological community conducted the Cooperative Convective Precipitation Experiment (CCOPE). The measurements collected during this project comprise the largest and most com-prehensive data set ever acquired in and around individual thunderstorms on the high plains of North America. The resultant archive contains approximately 300 billion bits of informa-tion compiled by state-of-the-art instrumentation in a field setting. The principal data systems utilized during CCOPC included 8 ground-based radars (7 of which had Doppler capability), 13 instrumented research aircraft, 6 sites from which balloon-borne instruments were launched, and a network of 123 surface stations. Our data processing goal has been to integrate all of these measurements into an accurate and com-plete three-dimensional description of any thunderstorm observed at any point throughout its history. Furthermore, this three-dimensional storm description must be embodied in a digi-tal structure that can be easily manipulated, altered, and displayed. Our presentation will focus on the procedures employed in reducing these diverse measurements to common spatial and temporal scales. The final product is a regularly spaced multi-dimensional Cartesian coordinate system at a discrete analysis time where each grid location contains the set of relevant meteorological parameters. A recently developed soft-ware package for analyzing the information in these data structures will also be discussed.

Presented is a discussion of the potential value of three-dimensional displays augmenting optical design software. Specifically, we discuss physical optics modeling methods and a configuration-based approach, which allows users to view the optical system in a three-dimensional projection. This approach utilizes elementary ray-trace capabilities used in providing geometric beam data to the physical optics part of the code. A global three-dimensional coordinate system provides a framework in which to locate components, and compute beam intercepts with the components in the optical systems.

This review paper summarizes the development and applications of computer techniques for the representation of three-dimensional data in the future flight station. It covers the development of the Lockheed-NASA Advanced Concepts Flight Station (ACFS) research simulators. These simulators contain: A Pilot's Desk Flight Station (PDFS) with five 13- inch diagonal, color, cathode ray tubes on the main instrument panel; a computer-generated day and night visual system; a six-degree-of-freedom motion base; and a computer complex. This paper reviews current research, development, and evaluation of easily modifiable display systems and software requirements for three-dimensional displays that may be developed for the PDFS. This includes the analysis and development of a 3-D representation of the entire flight profile. This 3-D flight path, or "Highway-in-the-Sky", will utilize motion and perspective cues to tightly couple the human responses of the pilot to the aircraft control systems. The use of custom logic, e.g., graphics engines, may provide the processing power and architecture required for 3-D computer-generated imagery (CGI) or visual scene simulation (VSS). Diffraction or holographic head-up displays (HUDs) will also be integrated into the ACFS simulator to permit research on the requirements and use of these "out-the-window" projection systems. Future research may include the retrieval of high-resolution, perspective view terrain maps which could then be overlaid with current weather information or other selectable cultural features.

It is desirable to extend the bene-fits of several new and emerging imag-ing modalities (such as computed tomography and nuclear magnetic reso-nance imaging) by providing means for visualizing an entire volume of image data at the same time. Such advances have been reported previously by us in the area of computer-generated holog-raphy, and in the generation of real-time imagery [1,2,3]. This paper reports on the applica-tion of these developed techniques to several diverse problems, including x-ray imaging of the pelvic girdle, neutron imaging of nuclear fuel rod bundles, and optical imaging of fluid flow in a volume. The particular problems and solutions involved in displaying these types of data are discussed.

Perspective transformation is one of many cues to 3d spatial information about objects in a 2d image. We can de-termine some 3d information such as the directions of lines or orientation of surfaces based on vanishing points(VP's) or vanishing lines(VL's) in the image. This note describes an analysis of errors in 3d measurements caused by errors in VP or VL detection. We will calculate the range of error in surface orientation when VP or VL location is determined within a certain error. First we will make clear the role that the location of the VP or VL plays in the 3d measurements. We will consider the computation of length of line segments and surface orientation. Then, we will calculate the range of the error of these 3d measurements as a function of VP or VL errors. Finally we will show how to use constraints based on real world knowledge of the relationships between surface orientations to increase the precision of the error analysis.

The civil engineering and architectural design requires a proper visual materialisation of the designed edifice, or highway, etc., and of its future environment. This can be done by a computer having a digitized drawing of the edifice and a digitized map of the environment as input. The output is a plotted perspective stereogram of the whole, that is a couple of perspectives which gives a three dimensional effect if viewed by means of a stereoscope. The paper gives the values of the perspective parameters taking into account the human physiology of stereoperception and the object-observer geometry. The paper says nothing new from the technological point of view: its object is to give an easy way for an effective three dimensional presentation, in a graphic two dimensional form, of civil engineering plans.

Two-dimensional data such as photos, X rays, various types of satellite images, sonar, radar, seismic plots, etc., in many cases must be analyzed using frame buffers for purposes of medical diagnoses, crop estimates, mineral exploration, and so forth. In many cases the same types of sensors used to gather such samples in two dimensions can gather 3D data for even more effective analysis. Just as 2D arrays of data can be analyzed using frame buffers, three-dimensional data can be analyzed using SOLIDS-BUFFEPmemories. Image processors deal with samples from two-dimensional arrays, and are based on frame buffers. The SOLIDS PROCESSOR system, deals with samples from a three-dimensional volume, or solid, and is based on a 3D frame buffer. This paper focuses upon the SOLIDS-BUFFER system, as used in the INSIGHT SOLIDS-PROCESSOR system from Phoenix Data Systems.