edu.mit.sketch.toolkit
Class SimpleClassifier

java.lang.Object
  |
  +--edu.mit.sketch.toolkit.Classifier
        |
        +--edu.mit.sketch.toolkit.SimpleClassifier

public class SimpleClassifier

extends Classifier

This class implements Classifier. The SimpleClassifier will deal with single stroke input only. There is a SimpleClassifier object associated with each stroke. In conjunction with this class is the StrokeData class. The users must creare a StrokeData object and pass it to this class.

Field Summary

double	dd_dt_average_scale This is how much we scale the average curvature to get the threshold for curvature fit.
Vertex[]	direction_fit Fit based on curvature
int	direction_window_width This is the number of points to be included for deriving direction.
Vertex[]	final_fit Best hybrid fit
int	fit_method This is the fit method to be used when deriving the direction.
Vertex[]	points Speed original points in the stroke.
double	polyline_vs_general_path_bias Final fit LSQE is scaled by this, and compared against the general path LSQE.
double	speed_average_scale This is how much we scale the average speed to get the threshold for the speed fit.
Vertex[]	speed_fit Speed fit
double	test_line_scale This controls the strictness of line classification.

Fields inherited from class edu.mit.sketch.toolkit.Classifier

COMPLEX, ELLIPSE, error_calculator, LINE, POLYGON, stroke_data

Constructor Summary

SimpleClassifier(StrokeData stroke_data)
Constructs a Classifier object.

Method Summary

int	classify() Returns the best fit type for the strokeData
int[]	classify(double error_bound) Returns the set of types that fit the stroke with error value less than the bound specified, (not necessarily sorted by goodness).
Approximation[]	classifyAndRank() Returns an array of approximations to the input stroke sorted by their goodness.
GeneralPath	getComplexApproximation() Returns the best fit GeneralPath for the StrokeData.
double	getComplexError() Returns the error between the best fit GeneralPath and the strokeData.
Ellipse	getEllipseApproximation() Returns the best fit Ellipse for the StrokeData.
double	getEllipseError() Returns the error between the best fit ellipse and the strokeData.
Line	getLineApproximation() Returns the best fit Line for the StrokeData.
double	getLineError() Returns the error between the best fit line and the strokeData.
Polygon	getPolygonApproximation() Returns the best fit Polygon for the StrokeData.
double	getPolygonError() Returns the error between the best fit polygon and the strokeData.
boolean	isComplex() Returns true if the best fit for the strokeData is a complex shape.
boolean	isComplex(double error_bound) Returns true if the best fit complex shape has error measure less than error_bound.
boolean	isEllipse() Returns true if the best fit for the strokeData is an Ellipse.
boolean	isEllipse(double error_bound) Returns true if the best fit ellipse for the strokeData is less than error_bound.
boolean	isLine() Returns true if the best fit shape for the strokeData is a Line.
boolean	isLine(double error_bound) Returns true if the best fit line has an error of less than error_bound.
boolean	isPolygon() Returns true if the best fit for strokeData is a polygon.
boolean	isPolygon(double error_bound) Returns true if the best fit polygon has an error of less than bound.

Methods inherited from class java.lang.Object

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

Field Detail

direction_window_width

public int direction_window_width

This is the number of points to be included for deriving direction.

fit_method

public int fit_method

This is the fit method to be used when deriving the direction.

dd_dt_average_scale

public double dd_dt_average_scale

This is how much we scale the average curvature to get the threshold for curvature fit. Making this too high of too low results in fewer points in the output fit.

speed_average_scale

public double speed_average_scale

This is how much we scale the average speed to get the threshold for the speed fit. Making this too high of too low results in fewer points in the output fit.

test_line_scale

public double test_line_scale

This controls the strictness of line classification. Higher threshold means less strict.

polyline_vs_general_path_bias

public double polyline_vs_general_path_bias

Final fit LSQE is scaled by this, and compared against the general path LSQE. Making this larger creates a bias towards complex fits.

final_fit

public Vertex[] final_fit

Best hybrid fit

speed_fit

public Vertex[] speed_fit

Speed fit

direction_fit

public Vertex[] direction_fit

Fit based on curvature

points

public Vertex[] points

Speed original points in the stroke.

Constructor Detail

SimpleClassifier

public SimpleClassifier(StrokeData stroke_data)

Constructs a Classifier object. You must pass it a StrokeData object. Contains a default error calculator, which is just the squared error.

Method Detail

classify

public int classify()

Returns the best fit type for the strokeData

Overrides:

classify in class Classifier

classify

public int[] classify(double error_bound)

Returns the set of types that fit the stroke with error value less than the bound specified, (not necessarily sorted by goodness).

Overrides:

classify in class Classifier

classifyAndRank

public Approximation[] classifyAndRank()

Returns an array of approximations to the input stroke sorted by their goodness.

isPolygon

public boolean isPolygon()

Returns true if the best fit for strokeData is a polygon.

Overrides:

isPolygon in class Classifier

isPolygon

public boolean isPolygon(double error_bound)

Returns true if the best fit polygon has an error of less than bound. The error is squared error

Overrides:

isPolygon in class Classifier

isEllipse

public boolean isEllipse()

Returns true if the best fit for the strokeData is an Ellipse.

Overrides:

isEllipse in class Classifier

isEllipse

public boolean isEllipse(double error_bound)

Returns true if the best fit ellipse for the strokeData is less than error_bound.

Overrides:

isEllipse in class Classifier

isComplex

public boolean isComplex()

Returns true if the best fit for the strokeData is a complex shape.

Overrides:

isComplex in class Classifier

isComplex

public boolean isComplex(double error_bound)

Returns true if the best fit complex shape has error measure less than error_bound.

Overrides:

isComplex in class Classifier

isLine

public boolean isLine()

Returns true if the best fit shape for the strokeData is a Line.

Overrides:

isLine in class Classifier

isLine

public boolean isLine(double error_bound)

Returns true if the best fit line has an error of less than error_bound.

Overrides:

isLine in class Classifier

getPolygonError

public double getPolygonError()

Returns the error between the best fit polygon and the strokeData. The error is calculated according to the ErrorCalculator, which uses squared error if not set by the user.

Overrides:

getPolygonError in class Classifier

getEllipseError

public double getEllipseError()

Returns the error between the best fit ellipse and the strokeData. The error is calculated according to the ErrorCalculator, which uses squared error if not set by the user.

Overrides:

getEllipseError in class Classifier

getComplexError

public double getComplexError()

Returns the error between the best fit GeneralPath and the strokeData. The error is calculated according to the ErrorCalculator, which uses squared error if not set by the user.

Overrides:

getComplexError in class Classifier

getLineError

public double getLineError()

Returns the error between the best fit line and the strokeData. The error is calculated according to the ErrorCalculator, which uses squared error if not set by the user.

Overrides:

getLineError in class Classifier

getPolygonApproximation

public Polygon getPolygonApproximation()

Returns the best fit Polygon for the StrokeData. I have used the class Polygon here because we are using a standard library of shapes which may or may not be the same as the Java defined shapes.

Overrides:

getPolygonApproximation in class Classifier

getEllipseApproximation

public Ellipse getEllipseApproximation()

Returns the best fit Ellipse for the StrokeData. I have used the class Ellipse here because we are using a standard library of shapes which may or may not be the same as the Java defined shapes.

Overrides:

getEllipseApproximation in class Classifier

getComplexApproximation

public GeneralPath getComplexApproximation()

Returns the best fit GeneralPath for the StrokeData. I have used the class GeneralPath here because we are using a standard library of shapes which may or may not be the same as the Java defined shapes.

Overrides:

getComplexApproximation in class Classifier

getLineApproximation

public Line getLineApproximation()

Returns the best fit Line for the StrokeData. I have used the class Line here because we are using a standard library of shapes which may or may not be the same as the Java defined shapes.

Overrides:

getLineApproximation in class Classifier