Package cz.vutbr.fit.layout.cormier.impl

Class LineDetector

java.lang.Object

cz.vutbr.fit.layout.cormier.impl.LineDetector

public class LineDetector
extends Object

Detector for semantically significant lines for the CormierSegmentation.

Author:

František Maštera

See Also:

Michael Cormier (2018). Computer Vision on Web Pages: A Study of Man-Made Images. UWSpace.

Constructor Summary

Constructors

Constructor	Description
LineDetector(int maxLineLength, float edgeProbabilityThreshold, int monteCarloTrials)

Method Summary

Modifier and Type	Method	Description
float	getEdgeProbabilityThreshold()
int	getMaxLineLength()
int	getMonteCarloTrials()
float	lineProbability(double[] edgeProbabilities)
void	setEdgeProbabilityThreshold(float edgeProbabilityThreshold)
void	setMaxLineLength(int maxLineLength)
void	setMonteCarloTrials(int monteCarloTrials)

Methods inherited from class java.lang.Object

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

Constructor Detail

LineDetector

public LineDetector(int maxLineLength,
                    float edgeProbabilityThreshold,
                    int monteCarloTrials)

Parameters:

maxLineLength - Maximum length of a segmentation line (in pixels) before the algorithm splits it in half and processes each half separately.

edgeProbabilityThreshold - Minimum probability of an edge in each pixel (obtained from EdgeDetector.getEdgeProbabilities(Mat)) required in order to be considered significant.

monteCarloTrials - Number of Monte Carlo trials for determining the probability that line is semantically significant.

Method Detail

getMaxLineLength

public int getMaxLineLength()

Returns:

Maximum length of a segmentation line (in pixels) before the algorithm splits it in half and processes each half separately.

setMaxLineLength

public void setMaxLineLength(int maxLineLength)

Parameters:

maxLineLength - Maximum length of a segmentation line (in pixels) before the algorithm splits it in half and processes each half separately.

getEdgeProbabilityThreshold

public float getEdgeProbabilityThreshold()

Returns:

Minimum probability of an edge in each pixel (obtained from EdgeDetector.getEdgeProbabilities(Mat)) required in order to be considered significant.

setEdgeProbabilityThreshold

public void setEdgeProbabilityThreshold(float edgeProbabilityThreshold)

Parameters:

edgeProbabilityThreshold - Minimum probability of an edge in each pixel (obtained from EdgeDetector.getEdgeProbabilities(Mat)) required in order to be considered significant.

getMonteCarloTrials

public int getMonteCarloTrials()

Returns:

Number of Monte Carlo trials for determining the probability that line is semantically significant.

setMonteCarloTrials

public void setMonteCarloTrials(int monteCarloTrials)

Parameters:

monteCarloTrials - Number of Monte Carlo trials for determining the probability that line is semantically significant.

lineProbability

public float lineProbability(double[] edgeProbabilities)

Parameters:

edgeProbabilities - Edge probabilities obtained from EdgeDetector.getEdgeProbabilities(Mat) for each pixel in the line to be checked in original order (i.e. not sorted).

Returns:

Probability that the line made of pixels with given edge probabilities is semantically significant for segmentation.