Package org.ddogleg.nn

Interface NearestNeighbor<P>

All Known Implementing Classes:
KdForestBbfNearestNeighbor, KdTreeNearestNeighbor, VpTree, WrapExhaustiveNeighbor
public interface NearestNeighbor<P>

Abstract interface for finding the nearest neighbor to a user specified point inside of a set of points in K-dimensional space. Solution can be exact or approximate, depending on the implementation. The distance metric is intentionally left undefined and is implementation dependent.

WARNING: Do not modify the input lists until after the NN search is no longer needed. If the input lists do need to be modified, then pass in a copy instead. This restriction reduced memory overhead significantly.

  • Nested Class Summary

    Nested Classes
    Modifier and Type
    Interface
    Description
    static interface 
    NearestNeighbor.Search<P>
    An independent search instance.

  • Method Summary

    Modifier and Type
    Method
    Description
    NearestNeighbor.Search<P>
    createSearch()
    Creates a new search for this data structure.
    void
    setPoints(List<P> points, boolean trackIndices)
    Specifies the set of points which are to be searched.

  • Method Details

    • setPoints

      void setPoints(List<P> points, boolean trackIndices)
      Specifies the set of points which are to be searched.
      Parameters:
      points - Set of points.
      trackIndices - If true it will keep track of the index. Making it easy to associate data.

    • createSearch

      NearestNeighbor.Search<P> createSearch()
      Creates a new search for this data structure. This is intended to enabled concurrent searches. After setPoints(java.util.List<P>, boolean) has been called and returned, each searched can be called independently in separate threads. Do not call setPoints(java.util.List<P>, boolean) which a search is being performed.
      Returns:
      A new search object for this instance.