Computer graphics processing and selective visual display system – Computer graphics processing – Three-dimension
Patent
1996-10-02
1999-07-13
Powell, Mark R.
Computer graphics processing and selective visual display system
Computer graphics processing
Three-dimension
345515, G06F 100
Patent
active
059233311
DESCRIPTION:
BRIEF SUMMARY
The present invention relates to a method of generation of computer-generated images using a spherical buffer.
It is applicable notably in the field of image processing by digital computers or graphic computers working with a database describing a scene in 3 or 4 dimensions (space and space/time). The computer images obtained using the invention can be utilized as distance images, for the calculation of the propagation of light energy or as final computer images to be viewed.
There are known methods of generation of computer images using the surface representation for solid objects or plane or hemispherical projection, i.e. the projection of the image on one or more planes or on a hemisphere. This projection is for example described in the article "The hemisphere radiosity method: a tale of 2 algorithms", Stephen H. Spencer, Review of Proceedings in Eurographics, 1991. This method consists in projecting onto a hemisphere centered about a viewpoint of the 3D scene and which is the point of observation, the selected polygons corresponding to this surface representation, also called "facets".
In addition, there are many known methods of determination of the hidden parts of a 3D scene. A simple method of determining the visible parts of the polygons to be projected, known as "Z-buffer" or "depth buffer", consists in calculating the depth or distance for each pixel in the projected image. If the polygon considered gives a value less than that stored in this "buffer", the old value is replaced by the new calculated value. This technique is utilized for plane projections, in other words when the sampling corresponding to the pixels of the projected image is made on one or several flat surfaces, for example on a plane located in the neighborhood of the point of projection in the case of the hemiplane algorithm, or on 5 faces of a hemicube in the case of the hemicube algorithm.
A determination of the visible parts of a 3D scene from a point of observation is described in the French patent application No. 93 04054 filed by the Applicant; it relates in particular to an algorithm that enables determination of the visible segments by hemispherical projection and calculation of the intersections of the polygons projected.
Image synthesis by plane projection encounters the problem of spectrum overlap known as "aliasing". This phenomenon causes defects in the image: "staircase"-like contours, blinking due to the random disappearance of small polygons, defective shadow contours, and so on which are due to the insufficient spatial and temporal sampling used.
This is in part due to the geometric distortions of projection during the transition from the 3D image to the 2D image resulting in non-uniform sampling for the 2D image. These distortions result for example from the projection of a 3D space onto a flat surface, a projection which does not respect the solid angles; strictly, for a given solid angle, the projection area should vary as a function of the direction of this solid angle in space. An adaptive sampling of the projection plane enables the aliasing phenomenon to be reduced. This sampling is such that the area of the pixels generated is not constant but corresponds to a substantially constant projected solid angle, whatever the direction. However, even when classic filtering is used, the quality of the results is limited, since they are based on approximations. Another solution consists in projecting the scene on a hemisphere, in which case the projection respects the solid angles, but the distortion then occurs in the later stage in which the projections of angles on a plane must be processed.
The problems of image synthesis are also due, at a more basic level, to the analog-to-digital conversion, which implies loss of information. This phenomenon is particularly present at the contours of the facets, at the boundary between adjacent facets and notably for the label information of the visible facet, since a value must be chosen among several for the corresponding pixel; for the two other types of information rela
REFERENCES:
patent: 4583185 (1986-04-01), Heartz
patent: 4715005 (1987-12-01), Heartz
patent: 4807158 (1989-02-01), Blanton et al.
patent: 4821212 (1989-04-01), Heartz
patent: 4855937 (1989-08-01), Heartz
patent: 4862388 (1989-08-01), Bunker
patent: 4928250 (1990-05-01), Greenberg et al.
patent: 5051734 (1991-09-01), Lake, Jr.
patent: 5222896 (1993-06-01), Smith, Jr.
patent: 5579456 (1996-11-01), Cosman
patent: 5694530 (1997-12-01), Goto
Spencer, Stephen N., "The Hemisphere Radiosity Method: A Tale of Two Algorithms", Eurographics Workshop on Photosimulation, Realism and Physics in Computer Graphics, May 21, 1990, pp. 129-135.
Regan, Matthew, et al, "A Low Latency Virtual Reality Display System", Technical Report No: 92/166, Manash University, Victoria, Australia, Sep. 1992, pp. 1-12.
Cohen, Michael, et al, "Radiosity and Realistic Image Synthesis" Academic Press professional, 1993, pp. 1-6, 65-107, 285-287 and 331-347.
Hearn, Donald, et al, "Computer Graphics", Prentis Hall, 1994, pp. 544-553.
Foley, James, et al, "Computer Graphics Principles and Practice", Addison-Wesley, 2ed, 1990, pp. 792-807.
Blonde Laurent
Dusseux Jean-Christophe
Powell Mark R.
Thomson Broadband Systems
LandOfFree
Method of generation of computer-generated images using a spheri does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of generation of computer-generated images using a spheri, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of generation of computer-generated images using a spheri will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2280517