Abstract:Direct volume rendering techniques use semi-transparency to visualize much more important features in the three-dimensional scalar sample volumes than isosurface rendering methods. However, since all voxels participate in the generation of each image,these techniques are computationally expensive and far from interactive manipulation. In fact, on the base of the boundary representation scheme for the volumetric data, the 3D spatial coherence can be fully exploited to drastically reduce the rendering time for arbitrary viewing orientation. This data model only considers the bounding cells pierced by theisosurfaces of specified materials ; therefore, it requires O(kn²) memory cells to represent O(n³) volumetric data, and produces a high degree of data compression. A PC-based rendering algorithm in use of the boundary scheme is proposed in this paper, which producesresults visually similar to the volume rendering techniques, but speedup at least an order of magnitude. Examples from medical CT studies are given in contrast to the results from ray casting method both in image quality and in processing time.

Abstract:Usually, a whole architectural model contains millions of polygons. Visibility preprocessing is necessary for real time walkthroughs in such an environment. When the viewpoint is in the building, only a small part of polygons can be seen by the viewer, the others are all hidden. The display will be speeded up if people just compute this small part of polygons. In this paper, an efficient algorithm of visibility preprocessing based on the "room" structure is proposed. During the real time walkthroughs, the visible polygons instead of those of the whole building are retrieved from the database and are sent to the hideor render pipeline.

Abstract:Virtual environment is a typical application of time-critical computing or time-critical rendering. This paper addresses the level of detail representing models in time-critical rendering. The authors present an automatic generation algorithm for models at multiple levels of detail based on triangle mesh simplification. The algorithm makes multiple passes over an existing triangle mesh, using local geometry and topology to remove venices that pass a distance or angle criterion. The holes left by the vertex removal are patched using a local triangulation process. Examples illustrate efficiency of the algorithm.

Abstract:A technique for painting directly on 3D surfaces is proposed to synthesize complex textures in this paper. Different from the conventional 2D paint system, the color painted by each stroke is applied as surface attributes to set up various textures and shading effects on 3D surfaces in real time. The approach allows users to continuously blend the texture on adjacent surface and blend multi-layer textures on the same surface.Results demonstrate that the method can be employed to create very complex textures.

Abstract:This paper proposes a new deformation control method based on the factor curves. Both Bars's global nonlinear deformation and Watt's space-time factor curves depend on some kinds of analytical expressions hence. The authors classify the factor curves and provide a unified deformation control method which is convenient for animation systems. By taking the factor curves as animation parameters, the new control method not only includes Bars's tapering and twisting operations but also Watt's factor curve control method. Animators can control the deformation effectively with the method.

Abstract:This paper presents an algorithm for computing quantities involving integration of NURBS curves, such as areas, volumes of revolution, first moments of area, centroids etc., of regions bounded by plane NURBS curves. For quadratic NURBS curves,closed form integral solutions are derived, thus all integral values can be directly obtaind by the coordinates of control points and weights without approximation error. For cubic NURBS curves, an approximation method is given, which provides a simple error bound and has a speed advantage for small tolerance.

Abstract:This paper presents a feature extraction and visualization method for time - dependent data field. Feature definition and attributes are discussed. Then it gives a border tracking method using two edge operators to track the border of feature area. The relationship of the feature is represented by Fourier descriptors. The implementation and problems of this method are also discussed.

Abstract:This paper presents a GIVE(general interactive visualization environment),which supports the development of ViSC applications. GIVE owns a data flow kernel, a visual programming interface and extensible module libraries. These provide users a convenient interactive module-level programming environment. GIVE has these features: it introduces control nodes, e. g. IF-THEN -ELSE, COUNT -LOOP, WHILE -LOOP,to support conditional alternative structure and loop structure for building complex ViSC application; it is an opened environment, the module libraries and data types are both extensible; the module libraries contain a lot of algorithms.

Abstract:This paper presents a distributed shared buffer paradigm based on message passing in distributed computing environment. According to this paradigm, this paper implements a parallel volume rendering algorithm. In order to improve the efficiency of the cache, the parallel algorithm take advantage of octree classification to increase local coherence of the cache in data space. In image space, the parallel algorithm use the Hilbert pixel traversal order to increase the temporal coherence of the cache. The experiments demonstrate good speedup and salability of this approach.

Abstract:This paper studies on visualization of fuzzy mutlivariate optimization in complicated engineering systems, and presents an algorithm about drawing and calculating fuzzy sub-domain. Then some mathematical analysis are made to these problems, the distribution law of fuzzy workable domain in design space is displayed intuitively.

Abstract:The FVR(frequency domain volume rendering) approach explores a 3D data set faster than traditional spatial domain methods because of its innate lower computation al complexity. Unfortunately, images produced by the FVR methods always lack occlusion or depth cue since it is hard to represent and utilize the depth information within frequency domain efficiently. Some researchers have sought to make a remedy of this by means such as of linear depth cuing. This approach, however, is still far different from the conventional spatial domain methods which, using the color composition technique, generate images with realistic depth effect. In this paper, the discrete color composition of spatial domain algorithms is reviewed, before an analytical result of the continuous color composition is deducted. These equations become feasible for calculating in frequency domain on certain assumptions such as the transparency inside the data set is distributed homogeneously. The authors continue to discuss the detail implementation of this new algorithm based on their previous work, the substance classification FVR approach. The experimental result shows that both the computational complexity and the run-time cost are kept at the same level of former FVR algorithms.