Inferring Quality In Point Cloud-Based Three-Dimensional Objects Using Topographical Data Analysis

Disclosed are various embodiments for inferring quality in point cloud-based three-dimensional objects using topographical data analysis. A first graph is generated representing a three-dimensional model, each vertex in the first graph representing a respective connected component within a layer of the three-dimensional model and each edge in the first graph representing a connection between two respective connected components within two respective layers of the three-dimensional model. A second graph representing negative space associated with the three-dimensional model is also generated, each vertex in the second graph representing a connected component of a negative space region within the layer of the three-dimensional model and each edge in the second graph representing a connection between two respective connected components with two respective layers of the three-dimensional model. A persistent homology analysis is applied to the first graph to determine whether a hole or tunnel exists in each vertex of the first graph. An error with the three-dimensional model can then be identified based at least in part on the first graph, the second graph, and the persistent homology analysis.

Paul Rosen. Inferring Quality In Point Cloud-Based Three-Dimensional Objects Using Topographical Data Analysis. Workshop on Data Systems for Interactive Analysis (DSIA) at IEEE VIS, 2021.

@misc{rosen2021inferring,
  title = {Inferring Quality in Point Cloud-based Three-Dimensional Objects using
    Topographical Data Analysis},
  author = {Rosen, Paul},
  year = {2021},
  note = {University of South Florida, Application~16/725,051, Filed December 2019,
    Publication Number US 10,977,861},
  abstract = {Disclosed are various embodiments for inferring quality in point
    cloud-based three-dimensional objects using topographical data analysis. A first graph
    is generated representing a three-dimensional model, each vertex in the first graph
    representing a respective connected component within a layer of the three-dimensional
    model and each edge in the first graph representing a connection between two respective
    connected components within two respective layers of the three-dimensional model. A
    second graph representing negative space associated with the three-dimensional model is
    also generated, each vertex in the second graph representing a connected component of a
    negative space region within the layer of the three-dimensional model and each edge in
    the second graph representing a connection between two respective connected components
    with two respective layers of the three-dimensional model. A persistent homology
    analysis is applied to the first graph to determine whether a hole or tunnel exists in
    each vertex of the first graph. An error with the three-dimensional model can then be
    identified based at least in part on the first graph, the second graph, and the
    persistent homology analysis.}
}