Skip to content
University of Maryland LibrariesDigital Repository at the University of Maryland
    • Login
    View Item 
    •   DRUM
    • Theses and Dissertations from UMD
    • UMD Theses and Dissertations
    • View Item
    •   DRUM
    • Theses and Dissertations from UMD
    • UMD Theses and Dissertations
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Spatial Decompositions for Geometric Interpolation and Efficient Rendering

    Thumbnail
    View/Open
    umi-umd-1795.pdf (773.6Kb)
    No. of downloads: 2393

    Date
    2004-08-05
    Author
    Atalay-Satoglu, Fatma Betul
    Advisor
    Mount, David M
    Metadata
    Show full item record
    Abstract
    Interpolation is fundamental in many applications that are based on multidimensional scalar or vector fields. In such applications, it is possible to sample points from the field, for example, through the numerical solution of some mathematical model. Because point sampling may be computationally intensive, it is desirable to store samples in a data structure and estimate the values of the field at intermediate points through interpolation. We present methods based on building dynamic spatial data structures in which the samples are computed on-demand, and adaptive strategies are used to avoid oversampling. We first show how to apply this approach to accelerate realistic rendering through ray-tracing. Ray-tracing can be formulated as a sampling and reconstruction problem, where rays in 3-space are modeled as points in a 4-dimensional parameter space. Sample rays are associated with various geometric attributes, which are then used in rendering. We collect and store a relatively sparse set of sampled rays, and use inexpensive interpolation methods to approximate the attribute values for other rays. We present two data structures: (1) the <i>ray interpolant tree (RI-tree)</i>, which is based on a kd-tree-like subdivision of space, and (2) the <i>simplex decomposition tree (SD-tree)</i>, which is based on a hierarchical regular simplicial mesh, and improves the functionality of the RI-tree by guaranteeing continuity. For compact storage as well as efficient neighbor computation in the mesh, we present a pointerless representation of the SD-tree. An essential element of this approach is the development of a location code that enables efficient access and navigation of the data structure. For this purpose we introduce a location code, called an LPTcode, that uniquely encodes the geometry of each simplex of the hierarchy. We present rules to compute the neighbors of a given simplex efficiently through the use of this code. We show how to traverse the associated tree and how to answer point location and interpolation queries. Our algorithms work in arbitrary dimensions. We also demonstrate the use of the SD-tree for rendering atmospheric effects. We present empirical evidence that our methods can produce renderings of good quality significantly faster than simple ray-tracing.
    URI
    http://hdl.handle.net/1903/1812
    Collections
    • Computer Science Theses and Dissertations
    • UMD Theses and Dissertations

    DRUM is brought to you by the University of Maryland Libraries
    University of Maryland, College Park, MD 20742-7011 (301)314-1328.
    Please send us your comments.
    Web Accessibility
     

     

    Browse

    All of DRUMCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister
    Pages
    About DRUMAbout Download Statistics

    DRUM is brought to you by the University of Maryland Libraries
    University of Maryland, College Park, MD 20742-7011 (301)314-1328.
    Please send us your comments.
    Web Accessibility