w2kmeans Namespace Reference


Detailed Description

The w2kmeans namespace contains classes for the K-Means segmentation, motion estimation, and cluster attribute estimation algorithm.

It also contains comparision algorithms' (cross-correlation) classes.


Classes

class  AlarmAction
struct  AlarmRegion
class  AlarmTableCreator
class  Alarm
 Algorithm main class, whose methods are invoked by various listeners in w2alarm_main.cc Interface code auto-generated by w2algcreator. More...
class  ObjectImageComparator
 Aligns the objects in the forecast to the verification field, reports on the alignment required, and then does an image comparision. More...
class  PolynomialFit
 Fits a polynomial to data points. More...
class  ScoreTrack
 Algorithm main class, whose methods are invoked by various listeners in w2scoretrack_main.cc Interface code auto-generated by w2algcreator. More...
class  AdvectionFilter
class  CachedReader
 Caches reading for the kmeans set of algorithms. More...
class  ClusterAssociator
 Associates clusters across time. More...
class  MultistageClusterAssociator
 Uses a multi-stage strategy to associate clusters: (1) Project clusters to expected location based on motion field (2) Associate unique matches within overlap radius (3) Associate long-lived clusters within motion vector bounds (4) Associate closest if within lower of overlap and motion vector (5) Else assign new ID. More...
struct  ClusterTableRegion
 Specifies a region for use by ClusterTableAction classes. More...
class  ClusterTableAction
 An action that will be performed on a cluster table input to create a new output grid. More...
class  GridOperator
 Base-class of cluster-table-action objects that use grids as input. More...
class  ColumnOperator
 Base-class of cluster-table-action objects that use columns computed by previous operators as input. More...
class  ClusterTableCreator
 This class builds the cluster table based on a bunch of fields. More...
class  ClusterTableInput
 Holds the input as it is being constructed by the appropriate action objects. More...
class  ClusterInputHistory
 The state is arranged chronologically (earliest first). More...
class  CressmanAdvector
 Uses Cressman interpolation instead of "GET" to fill in missing pixels. More...
class  DefaultGridFilter
 A convenience class that provides default, do-nothing implementations of all the filter types. More...
class  ForecastMontage
 Creates a montage product of forecasts (different time intervals). More...
class  HierarchicalCluster
 The HierarchicalCluster is the final clustering result at a particular scale. More...
class  HistogramEqualizer
 This filter takes the input image and maps its data values such that the kmeans algorithm can work on it. More...
class  HSMotionEstimator
 Estimates motion using the Horn Schunk tracking equation. More...
class  ImageMotionEstimator
 Estimates motion using a constant sized neighborhood template at all valid pixels in the given image and using the minimum mean absolute error. More...
class  InversionFilter
 This filter takes the input image and maps its data values such that the kmeans algorithm can work on it. More...
class  KLTMotionEstimator
 Estimates motion using the Kanade-Lucas-Tomasi tracking equation. More...
class  KMeansSegmenter
 A hierarchical K-means segmentation algorithm that aggregates regions at different scales based purely on size. More...
class  MotionAdvector
 Advects given images given the motion estimates. More...
class  MotionEstimateHandler
 An interface for a motion estimator to communicate its estimates to a motion advector, for example. More...
class  MotionEstimateSupplier
 Base class for MotionEstimators -- subclasses include the abstract base class MotionEstimator which actually does the job and the MotionEstimateListener which gets its estimate by being attached to one or more URLs. More...
class  MotionEstimateListener
 Can be attached to an Index so that it listens for the motion estimates and provides it to attached MotionEstimateHandlers. More...
class  MotionEstimator
 Abstract base class of MotionEstimators that use successive inputs of images to form an estimate of the motion field. More...
class  MultiscaleSegmenter
 Abstract base class that defines the interface of multiscale segmenters. More...
struct  Pixel2D
struct  PixelRow
class  ProbabilisticAdvector
 Uses the motion vectors to advect the data forward, but in a probabilistic sense, so that the output images are not forecasts of intensity, but of the probability of location. More...
class  ProbabilisticSwathAdvector
 Uses the motion vectors to advect the data forward, but in a probabilistic sense, so that the output images are a swathe of the probability of location until (not at) the desired time. More...
class  PutGetAdvector
 Uses the motion vectors to advect the data forward, then fills in any holes with a "GET" to grab the data from the motion estimate at the current point. More...
class  ScaleCalculator
class  SegmentMotionEstimator
 Estimates motion using a template with the shape of a Segment (or cluster, obtained from a Segmenter) such that each Segment/cluster gets a motion estimate. More...
class  SwathUpdater
 Updates swath grid at a single grid point. More...
class  SwathMaxUpdater
 Keeps maximum value at a single grid point. More...
class  SwathRateUpdater
 Accumulates value at a single grid point. More...
class  SwathAdvector
 Uses the motion vectors to advect the data forward, but fills in complete path to final location, so that a swath is created. More...
class  WeightingScheme
 This class does the calculations for the spatial interpolation performed by SegmentMotionEstimator. More...


Generated on Fri May 4 13:40:24 2012 for WDSS-IIw2algs by  doxygen 1.4.7