SimulationViewerSurpCont Class Reference

This class provides an interface into SurpriseControl (ASAC). More...

#include <Neuro/SimulationViewerSurpCont.H>

Inheritance diagram for SimulationViewerSurpCont:

[legend]

Collaboration diagram for SimulationViewerSurpCont:

[legend]

List of all members.

Public Member Functions
void	init (const ushort baseSizeX, const ushort baseSizeY)
	Initialize everything. Input the raw image size.
Constructors and destructors

	SimulationViewerSurpCont (OptionManager &mgr, const std::string &descrName="ASAC Surprise Control", const std::string &tagName="SurpriseControl")
	Constructor. See ModelComponent.H.
virtual	~SimulationViewerSurpCont ()
	Destructor.
virtual void	setBrain (Brain *brain)
	Set our brain.
Protected Member Functions
	SIMCALLBACK_DECLARE (SimulationViewerSurpCont, SimEventInputFrame)
	Callback for when a new input frame is available.
	SIMCALLBACK_DECLARE (SimulationViewerSurpCont, SimEventClockTick)
	Callback for every clock tick.
	SIMCALLBACK_DECLARE (SimulationViewerSurpCont, SimEventSaveOutput)
	Callback for every time we should save our outputs.
virtual void	save1 (const ModelComponentSaveInfo &sinfo)
	Save our various results.
void	saveResults (const nub::ref< FrameOstream > &ofs)
	save results
Image< PixRGB< float > >	getResult ()
	return the surprise controlled image
std::vector< Image< PixRGB < float > > >	getDiffImages ()
	return the difference images if needed
std::vector< Image< float > >	getBetaImages ()
	return the beta images if needed
void	drawTime (Image< PixRGB< byte > > &image)
	Draw current time onto given image.
Protected Attributes
nub::ref< SpatialMetrics >	itsMetrics
	metrics that depend on the input size:
OModelParam< bool >	itsDrawDiffParts
	Should we output the difference image between the pre/post image.
OModelParam< bool >	itsDrawBetaParts
	should we output the surprise maps used here?
OModelParam< bool >	itsDrawBiasParts
	should we output the bias maps used here?
OModelParam< bool >	itsDrawSeperableParts
	should we output the seperable filter layers used here?
OModelParam< std::string >	itsConfigFile
	config file to open
OModelParam< LevelSpec >	itsLevelSpec
	LevelSpec used by our channels, used to compute output dims.

Detailed Description

This class provides an interface into SurpriseControl (ASAC).

Surprise Control takes in an image and a set of final stage conspicuity images at different scales and uses them to drive different filters that are designed to reduce/control surprise in an image.

In general, any type of conspicuity map can be used, but it's primarily designed to work with Surprise images. However, this version requires the standard default scales and channels (rg,by... etc). It does not know how to handle different features such as junctions.

You need to supply:

A FrameSeries of raw images one at a time (per usual)
A configFile , there should be a default in etc
A brain , this is needed to extract the conspicuity maps

This class will return:

A surprise controlled image
A difference image showing what it changed
A beta image showing its internal temporal smoothed surprise maps

T. Nathan Mundhenk <mundhenk@usc.edu>

Definition at line 84 of file SimulationViewerSurpCont.H.

Constructor & Destructor Documentation

SimulationViewerSurpCont::SimulationViewerSurpCont	(	OptionManager &	mgr,
		const std::string &	descrName = `"ASAC Surprise Control"`,
		const std::string &	tagName = `"SurpriseControl"`
	)