Package edu.gsu.cs.dmlab.tracking

Class PoissonObsModel

java.lang.Object

edu.gsu.cs.dmlab.tracking.PoissonObsModel

All Implemented Interfaces:

ISTObsModel

public class PoissonObsModel
extends Object
implements ISTObsModel

This class is used to calculate the observation cost of a given track using a Poisson model of observation that uses the change in the number of events detected in the current frame from the previous. It also uses the average from a number of pervious frames.

Author:

Dustin Kempton, Data Mining Lab, Georgia State University

Constructor Summary

Constructors

Constructor	Description
PoissonObsModel(ISTTrackingEventIndexer evntsIdxr, int span)	Constructor

Method Summary

Modifier and Type	Method	Description
void	finalize()
double	getObsProb(IBaseTemporalObject ev)	Returns the probability of an event being a true observation based on some model of observation.

Methods inherited from class java.lang.Object

clone, equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

PoissonObsModel

public PoissonObsModel(ISTTrackingEventIndexer evntsIdxr,
                       int span)

Constructor

Parameters:

evntsIdxr - Indexer which is used to get the expected change per frame.

span - The time span, in seconds, to calculate the expected change per frame on. (probably spanOf event type times the number of frames over which to calculate).

Method Detail

finalize

public void finalize()
              throws Throwable

Overrides:

finalize in class Object

Throws:

Throwable

getObsProb

public double getObsProb(IBaseTemporalObject ev)

Description copied from interface: ISTObsModel

Returns the probability of an event being a true observation based on some model of observation.

Specified by:

getObsProb in interface ISTObsModel

Parameters:

ev - The event to calculate the observation probability on.

Returns:

The probability of the event being a true detection with range [0,1].