Package edu.gsu.cs.dmlab.tracking.stages

Class EdgeWeightCalculatorStageTwo

  • All Implemented Interfaces:
    ISTEdgeWeightCalculator
    public class EdgeWeightCalculatorStageTwo
extends BaseEdgeWeightCalculator
    This class calculates the edge weights in the graph association problem. It is used for stage two as the edge between two tracks is calculated differently than in stages three and above. Namely, it doesn't use the motion model because it is assumed that track fragments are not long enough yet to be able to calculate motion on them. An implementation of the edge weight calculation based upon Kempton et. al 2015 and 2018.
    Author:
    Dustin Kempton, Data Mining Lab, Georgia State University

    • Constructor Detail

      • EdgeWeightCalculatorStageTwo

        public EdgeWeightCalculatorStageTwo​(ISTLocationProbCal enterLocProbCalc,
                                    ISTLocationProbCal exitLocProbCalc,
                                    ISTObsModel obsModel,
                                    ISTAppearanceModel appearanceModel,
                                    ISTFrameSkipModel skipModel)
        Constructor
        Parameters:
        enterLocProbCalc - The model used for calculating enter probability on a track.
        exitLocProbCalc - The model used for calculating exit probability on a track.
        obsModel - The model used for calculating the probability of a track being plausible.
        appearanceModel - The model used for calculating the probability of two tracks being linked base on their visual similarity.
        skipModel - The model used for calculating the probability that a track skipped n frames before being detected again.

      • EdgeWeightCalculatorStageTwo

        public EdgeWeightCalculatorStageTwo​(ISTLocationProbCal enterLocProbCalc,
                                    ISTLocationProbCal exitLocProbCalc,
                                    ISTObsModel obsModel,
                                    ISTAppearanceModel appearanceModel,
                                    ISTFrameSkipModel skipModel,
                                    double entExitMult,
                                    double obsMult,
                                    double assocMult)
        Constructor
        Parameters:
        enterLocProbCalc - The model used for calculating enter probability on a track.
        exitLocProbCalc - The model used for calculating exit probability on a track.
        obsModel - The model used for calculating the probability of a track being plausible.
        appearanceModel - The model used for calculating the probability of two tracks being linked base on their visual similarity.
        skipModel - The model used for calculating the probability that a track skipped n frames before being detected again.
        entExitMult - The multiplier value used to weight the enter/exit edges values.
        obsMult - The multiplier value used to weight the observation edge values.
        assocMult - The multiplier value used to weight the association edge values.

      • EdgeWeightCalculatorStageTwo

        public EdgeWeightCalculatorStageTwo​(ISTLocationProbCal enterLocProbCalc,
                                    ISTLocationProbCal exitLocProbCalc,
                                    ISTObsModel obsModel,
                                    ISTAppearanceModel appearanceModel,
                                    ISTFrameSkipModel skipModel,
                                    double entExitMult,
                                    double obsMult,
                                    double assocMult,
                                    double appearWeight,
                                    double skipWeight)
        Constructor
        Parameters:
        enterLocProbCalc - The model used for calculating enter probability on a track.
        exitLocProbCalc - The model used for calculating exit probability on a track.
        obsModel - The model used for calculating the probability of a track being plausible.
        appearanceModel - The model used for calculating the probability of two tracks being linked base on their visual similarity.
        skipModel - The model used for calculating the probability that a track skipped n frames before being detected again.
        entExitMult - The multiplier value used to weight the enter/exit edges values.
        obsMult - The multiplier value used to weight the observation edge values.
        assocMult - The multiplier value used to weight the association edge values.
        appearWeight - The weight for the appearance model values in the association edge values.
        skipWeight - The weight for the skip model values in the association edge values.