• Login
    View Item 
    •   JScholarship Home
    • Theses and Dissertations, Electronic (ETDs)
    • ETD -- Doctoral Dissertations
    • View Item
    •   JScholarship Home
    • Theses and Dissertations, Electronic (ETDs)
    • ETD -- Doctoral Dissertations
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Random Graph Modeling and Discovery

    Thumbnail
    View/Open
    HALLONQUIST-DISSERTATION-2015.pdf (21.95Mb)
    Date
    2015-10-16
    Author
    Hallonquist, Neil F.
    Metadata
    Show full item record
    Abstract
    In the first part of this thesis, we present a general class of models for random graphs that is applicable to a broad range of problems, including those in which graphs have complicated edge structures. These models need not be conditioned on a fixed number of vertices, as is often the case in the literature, and can be used for problems in which graphs have attributes associated with their vertices and edges. To allow structure in these models, a framework analogous to graphical models is developed for random graphs. In the second part of this thesis, we consider the situation in which there is an unknown graph that one wants to determine. This is a common occurrence since, in general, entities in the world are not directly observable, but must be inferred from some signal. We consider a general framework for uncovering these unknown graphs by a sequence of ‘tests’ or ‘questions’. We refer to this framework as graph discovery. In the third part of this thesis, we apply graph discovery to a problem in computer vision. To evaluate how well vision systems perform, their interpretations of imagery must be compared to the true ones. Often, image interpretations can be expressed as graphs; for example, vertices can represent objects and edges can represent relationships between objects. Thus, an image, before it is interpreted, corresponds to an unknown graph, and the interpretation of an image corresponds to graph discovery. In this work, we are interested in the evaluation of vision systems when these representation graphs are complex. We propose a visual Turing test for this purpose.
    URI
    http://jhir.library.jhu.edu/handle/1774.2/39624
    Collections
    • ETD -- Doctoral Dissertations

    DSpace software copyright © 2002-2016  DuraSpace
    Policies | Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of JScholarshipCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister

    DSpace software copyright © 2002-2016  DuraSpace
    Policies | Contact Us | Send Feedback
    Theme by 
    Atmire NV