The rendering of photorealistic rain has been previously studied for monoscopic viewing. We extend the monoscopic statistical rain models to simulate the behavior and distribution of falling rain for stereo viewing. Our goal is to be able to render in real-time frame rates. In this investigation we ignore the complex issues of scene illumination and concentrate on the parameters that produce a realistic rain distribution. Using the concept of retinal persistence we render a visible falling raindrop as a linear streak. To speed rendering we use pre-computed images of such rain streaks. Rain streak positions for the left- and right-eye views are created by generating random numbers that depend on the view volume of the scene. We permit interactive but controlled modification of rain parameters such as density and wind gusts. We compare our approach to the use of existing 2D-3D conversion methods. The results demonstrate that using commercial 2D-3D converters are not sufficient in producing realistic stereo rain effects. Future research will concentrate on including complex lighting interactions.
A new computation technique is presented for calculating pixel colors in anaglyph images. The method depends
upon knowing the RGB spectral distributions of the display device and the transmission functions of the filters in the
viewing glasses. It requires the solution of a nonlinear least-squares program for each pixel in a stereo pair and is based
on minimizing color distances in the CIEL*a*b* uniform color space. The method is compared with several techniques
for computing anaglyphs including approximation in CIE space using the Euclidean and Uniform metrics, the Photoshop
method and its variants, and a method proposed by Peter Wimmer. We also discuss the methods of desaturation and
gamma correction for reducing retinal rivalry.
We propose a method for real-time photorealistic stereo rendering of the natural phenomenon of fire. Applications
include the use of virtual reality in fire fighting, military training, and entertainment. Rendering fire in real-time
presents a challenge because of the transparency and non-static fluid-like behavior of fire. It is well known that, in
general, methods that are effective for monoscopic rendering are not necessarily easily extended to stereo rendering
because monoscopic methods often do not provide the depth information necessary to produce the parallax required for
binocular disparity in stereoscopic rendering. We investigate the existing techniques used for monoscopic rendering of
fire and discuss their suitability for extension to real-time stereo rendering. Methods include the use of precomputed
textures, dynamic generation of textures, and rendering models resulting from the approximation of solutions of fluid
dynamics equations through the use of ray-tracing algorithms. We have found that in order to attain real-time frame
rates, our method based on billboarding is effective. Slicing is used to simulate depth. Texture mapping or 2D images
are mapped onto polygons and alpha blending is used to treat transparency. We can use video recordings or prerendered
high-quality images of fire as textures to attain photorealistic stereo.
We evaluate a new method for computing color anaglyphs based on uniform approximation in CIE color space. The method depends on the spectral distribution properties of the primaries of the monitor and the transmission functions of the filters in the viewing glasses. We will compare the result of this method with several other methods that have been proposed for computing anaglyphs. To compute the color at a given pixel in the anaglyph image requires solving a linear program. We exploit computational properties of the simplex algorithm to reduce computation time by 72 to 89 percent. After computing the color at a pixel, a depth-first search is performed to collect all neighboring pixels with similar color so that a simple matrix-vector multiplication can be applied. We also parallelize the algorithm and implement it on a cluster environment. We discuss the effects of different data dividing schemes.
Gaseous phenomena such as clouds, fog, and mist have been difficult to render in realistic monoscopic imaging environments. Such phenomena are transparent, cast shadows, have dynamic behavior, and are of variable density. This paper describes a method based on splatting, billboarding, and alpha-blending that works well in a realistic real-time stereo imaging environment. Splatting is used to reconstruct a discretely sampled 3D volume to produce a 2D image with appropriate density. Efficient reconstruction is gained through the use of texture-mapped billboards whose transparencies are determined by a Gaussian reconstruction kernel. To achieve the fastest rendering, it is possible to orient all billboards to face the viewplane rather than the viewpoint. The parallax error introduced by this approach is analyzed. The authors give examples to illustrate how the number, position, and size of the billboards in a scene can be used to create different effects. The discussion does not treat the problems of self-shadowing or dynamic behavior although the approach could be used as the basis for simulating both.
Distance learning and virtual laboratory applications have motivated the use of inexpensive visual stereo solutions for computer displays. The anaglyph method is such a solution. Several techniques have been proposed for the production of anaglyphs. We discuss three approaches: the Photoshop algorithm and its variants, the least squares algorithm proposed by Eric Dubois that optimizes in the CIE color space, and the midpoint algorithm that minimizes the sum of the distances between the anagylph color and the left and right eye colors in CIEL*a*b*. Our results show that each method has its advantages and disadvantages in faithful color representation and in stereo quality as it relates to region merging and ghosting.
KEYWORDS: Vegetation, 3D modeling, Visualization, Eye, Eye models, 3D image processing, Transparency, Motion models, Visual process modeling, Systems modeling
Rendering of detailed vegetation for real-time applications has always been difficult because of the high polygon count in 3D models. Generating correctly warped images for nonplanar projection surfaces often requires even higher degrees of tessellation. Generating left and right eye views for stereo would further reduce the frame rate since information for one eye view cannot be used to redraw the vegetation for the other eye view. We describe an image based rendering approach that is a modification fo an algorithm for monoscopic rendering of vegetation proposed by Aleks Jakulin. The Jakulin algorithm pre-renders vegetation models from 6 viewpoints; rendering from an arbitrary viewpoint is achieved by compositing the nearest two slicings. Slices are alpha blended as the user changes viewing positions. The blending produces visual artifacts that are not distracting in a monoscopic environment but are very distracting in a stereo environment. We have modified the algorithm so it displays all pre-rendered images simultaneously and slicings are partitioned and rendered in a back-to-front order. This approach improves the quality of the stereo, maintains the basic appearance of the vegetation and reduces visual artifacts but it increases rendering time slightly and produces a rendering that is not totally faithful to the original vegetation model.
A flight simulator was developed for studying the behavior of pilots in power-off aircraft landing situations. The simulation environment includes a 5-meter hemispherical dome in which the authors have installed a cockpit from a Cessna aircraft. The dome manufacturers provided their version of OPEN GL 1.1. The graphics rendering software has undergone constant modification because of computer and projection hardware changes and a lack of knowledge and understanding of the manufacturer's undocumented version of OPEN GL. The development team was led to believe that real time rendering of photographic quality images from 3D models was possible using the existing hardware and software. This was not true even for very simple environments. Flat surfaces must undergo major tessellation to project correctly on a hemispherical dome. The number of polygons to be rendered is increased by orders of magnitude. The tessellation also reverses some depth relationships, which causes parts of objects to disappear and reappear during the simulation. In addition, aliasing artifacts are severe because of the limited resolution and lack of antialiasing capabilities of the hardware. The authors document their experiences and their solutions to some of the rendering problems encountered.
There are some interesting cases where image processing techniques applied independently to the left and right eye images can preserve some parallax within the accepted limits of vertical disparity. We investigate the application of simple geometric techniques to determine which ones alter stereo and to what extent they alter stereo. We consider the simple affine transformations such as translation, rotation, scale and shear and show that interesting results can be obtained in some cases by applying a transformation to one eye image and not the other. We show that any 2D transformation applied to one eye view of a stereo pair that preserves lines/planes must be a composition of x-shears, x-translations and x-scales.
KEYWORDS: 3D modeling, 3D acquisition, Light sources and illumination, 3D image processing, Optical spheres, Eye, Detection and tracking algorithms, Human-machine interfaces, 3D applications, Distortion
There are several techniques that can be used to produce morphs of 3D objects. The traditional solution is to apply 3D algorithms that transform the shape and attributes of one object into those of another. The problems in 3D morphing include avoiding self-intersections during the morph, specification of corresponding regions in the source and target objects and the imposition of geometric constraints on the objects. At first glance, the application of well understood 2D morphic techniques to stereo imags would seem to be reasonable and much simpler alternative to the production of 3D models and the application of 3D morphing to those modes. While it is true that in certain cases the applicant of 2D linear morphing techniques to stereo images produces effective morphs, the use of this technique places very strict geometric constraints on the objects being morphed. When linear 2D morphic technique are applied to stereo images where the parallax encoded in the images is of utmost importance, they linearly interpolate points between the source and target images which interpolates the parallax, also. We examine the ramifications of this limitation and discus the geometric constraints under which stereo morphing is useful.
Several authors have recently investigated the ability to compute intermediate views of a scene using given 2D images from arbitrary camera positions. The methods fall under the topic of image based rendering. In the case we give here, linear morphing between two parallel views of a scene produces intermediate views that would have been produced by parallel movement of a camera. Hence, the technique produces images computed in a way that is consistent with the standard off-axis perspective projection method for computing stereo pairs. Using available commercial 2D morphing software, linear morphing can be used to produce stereo pairs from a single image with bilateral symmetry such as a human face. In our case, the second image is produced by horizontal reflection. We describe morphing and show how it can be used to provide stereo pairs from single images.
Kaleidoscopes are normally constructed of three mirrors in a triangular pattern set in a tube. A changing 2D image is set at one end of the tube and observed by a monoscopic viewer at the opposite end. The orientation of the mirrors produces an infinite wall paper pattern with symmetries described by the algebraic structure known as the Dihedral Group D3. We show that the kaleidoscope can be used to generate 3D textures in a natural way. We generalize the kaleidoscope to allow binocular viewing, any number of mirrors, warped mirrors and objects which can move in and out of the tube at various depths. The images are produced using the rendering technique of ray tracing.
Present tracking schemes for virtual reality position sensing have a variety of features that make their use in applications such as large classrooms or remote locations difficult. A more natural tracking method would be a lightweight, low cost, and accurate inertial tracking system. Some commercial inertial systems are discussed. As a low cost alternative, a mouse based head self-tracker has been built at North Carolina State University. Its design and operational ideas are being extended to build a less cumbersome head tracker based on the rotational axes.
Translating rectangular 2D images in planes perpendicular to the line of sight can create image distortion known as keystoning or projection warp. This can occur when viewing several images on a plane from a fixed position, when cross viewing, or when using non-parallel mirrors to view stereo pairs. In the latter case, vertical disparity can become a problem. The projection warp and its inverse are derived as functions of the viewing and translation parameters. It is shown that vertical disparity can be significant, even in small image translations. The inverse transformation applied to the original image corrects the keystoning caused by the translation. Rendering of the inverse image is also discussed.
A stereoscopic drawing program is described which permits the user to display and manipulate quadric surfaces. The quadric surfaces are the 3-dimensional relatives of the ellipse, parabola and hyperbola and include ellipsoids, hyperboloids of one sheet, hyperboloids of two sheets, elliptic cones, elliptic paraboloids, and hyperbolic paraboloids. These surfaces have both implicit and parametric representations. A 3-button mouse is used to create and manipulate the surfaces. Rubber-banding can be used to define the surface and three dimensional transformations of the surface including scaling, rotation and translation are defined by mouse movement. A goal is to maintain a consistent and intuitive method of control for these surfaces, using techniques similar to those used in 2-dimensional drawing systems. The tessellation, color and shading characteristics of a surface can be determined interactively by the user.
We investigate the use of stereo in creating and manipulating ration Bezier tensor product surfaces. The application uses operating system provided 2D cursors and menuing which were found to be inferior to their stereo counterparts. Stereo manipulation was accomplished using a 3-button mouse which was determined to be an adequate input device. The interface is described and possible changes and additions are suggested.
We present a report on the effects of compressing stereo pairs using lossy compression techniques. We discuss the general functionality of all lossy compressors with particular emphasis on the CCITT/ISO JPEG Still Picture Compression Standard. We show that the application of lossy compressors to stereo pairs can cause ambiguities in the resulting stereo image and we develop a measure to quantify those ambiguities. We describe two new techniques which each help to decrease some of these anomalies. We also discuss the compression rates on some test images with and without the use of these new techniques.
Computer animation consists of a series of images each taken for a specific point in time. This can create temporal aliasing; motion, especially fast motion, appears discontinuous. Motion blur can be used to prevent a strobing effect and to enhance the perception of motion. Motion blur is commonly achieved by one of the following three methods: stretching the object along the path of motion, stochastic sampling in the time domain, and supersampling over time. When creating stereo images while using stochastic sampling, each eye sees different points on the object at a pixel location over time, creating discrepancies in the left and right views. Supersampling produces images that are averages of several disparate images of the same object, making it difficult to fuse the views. Traditional methods for creating motion blur can therefore produce images with ambiguous depth when combined with stereo. The stretched object method works well with stereo as there is consistency in both views. However, it exaggerates sizes of moving objects and lacks the blurring effect of the sampling methods. Our technique for creating motion blur in stereo is to stretch moving objects and then apply functions to give blurring effects such as transparency or fading.
We assume a left handed coordinate system with interocular distance e and viewer distance to the stereo window d < 0. Using the standard off-axis perspective projection to compute homologous points, the horizontal parallax of a point with coordinates (x, y, z) becomes ez/(z - d). Given a finite set of points D equals {(xj, yj, zj), 1 j in a z-buffer after applying hidden surface elimination. In the first case the solution is based on the root of a quadratic equation. In the second case we approximate the translated parallax function zj + v)/(zj + v - d) by the first term in its series expansion, (zj + v)/d , and use it to analyze the behavior of the sum near the minimum. If we suitably restrict the distance between the closet point and furthest point from the viewer, we argue that the minimum occurs to the right of max {d-zi} and at a root -zj of a parallax function. The root can be located by determining the index i where the function changes sign and evaluating the absolute parallax sum at values -zj where j is close to i.
KEYWORDS: Modulation, Spatial light modulators, 3D image processing, Medical imaging, Atmospheric optics, Lithium, Spatial filters, 3D displays, Lenses, Imaging systems
A spatial modulation display is described that permits the observation of a phantom or transparent image by several persons simultaneously and is suitable for medical imaging. The display uses spatial light modulators and large format convex lenses within a Schlieren optical system. The number of sectional images in a three-dimensional image is limited by the number of spatial light modulators. The display is electro-optical and requires no moving parts.
An interactive stereo system was implemented to allow a user to draw and modify B-Spline space curves. A time-multiplexed, liquid crystal shutter system was used to produce true three-dimensional output. A six degree-of-freedom input device was used to drive and control the stereoscopic cursor. Several different cursor types were investigated to determine which were suited to this application. Cursor characteristics were examined to determine their influence on the user''s ability to accurately and proficiently use the system.
A method for simultaneous generation of stereoscopic images of implicitly defined surfaces is presented. Ray tracing, a computationally intensive but `photo-realistic'' rendering technique, is used to generate these images. Traditional ray tracing of stereoscopic images involves generating left eye and right eye images separately, which effectively doubles the computational work involved. A method originally developed for animation and later adopted for stereo ray tracing achieves computational speedup in ray tracing stereoscopic images by generating the right eye image as the left eye image is generated. The work presented here extends that method to obtain stereoscopic images of implicitly defined surfaces of the form f(x,y,z) equals 0. Implicitly defined quadric surfaces such as spheres, ellipsoids, hyperboloids of one sheet and two sheets, and surfaces derived from fault-tolerant software modeling applications were rendered using this method. The speedups achieved for these surfaces under different stereoscopic setups are compared and tabulated.
With recent advances in 3-D technology, computer users have the opportunity to work within a natural 3-D environment; a flat panel LCD computer display of this type, the DTI-100M made by Dimension Technologies, Inc., recently went on the market. In a joint venture between DTI and NCSU, an object-oriented 3-D drawing application, 3-D Draw, was developed to address some issues of human interface design for interactive stereo drawing applications. The focus of this paper is to determine some of the procedures a user would naturally expect to follow while working within a true 3-D environment. The paper discusses (1) the interface between the Macintosh II and DTI-100M during implementation of 3-D Draw, including stereo cursor development and presentation of current 2-D systems, with an additional `depth'' parameter, in the 3-D world, (2) problems in general for human interface into the 3-D environment, and (3) necessary functions and/or problems in developing future stereoscopic 3-D operating systems/tools.
A report on the effects of quantizing stereo pairs computed for stereopsis is presented. Two methods for quantization of color images -- the octree method, which is elegant and fast, and the MaxMin method, which was designed for pseudo-random dithering -- are compared. Their application to stereo pairs, which have been computed using true color (24 bit), can cause anomalies in the resulting stereo image such as loss of depth, contradictory features leading to binocular rivalry, image discontinuities, and loss of definition. Techniques for image sampling to decrease processing time and color lookup table construction to minimize the anomalies are also discussed. A solution for quantizing stereo animations is suggested.
Previous researchers have applied vertical motion parallax using video cameras; these attempts were published under the VISIDEP trademark. Alternatives are examined that can exploit computer graphics. Such techniques include choosing different locations for the axis of rotation and using multiple images versus dual images to make the motion smoother. These techniques as well as others are described and compared for effectiveness in enhancing depth.
We examine the effect of using rotations for generation of the left- and right-eye perspective views of a stereoscopic image. We show that this approach to stereoscopic display of perspective views results in vertical parallax between the left- and right-eye views and we present an analytic expression that characterizes this parallax in terms of the center of rotation, the location of the image plane, and the angle of rotation. We also derive an analytic expression that shows that the rotation of the
perspective views results in a semicylindrical stereo window with center at approximately (0, 0, P12) and radius R/2, where R is the distance from the center of rotation to the center of projection. When this semicylindrical window is mapped to a flat display surface, relative depth relationships can be distorted.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.