The class binarize a grayvalue image. The segmentation algorithm is based on "Binarization of Historical Document Images Using Local Maximum and Minimum", Bolan Su, Shijian Lu and Chew Lim Tan, DAS 2010. In contrast to DkSegmentationSu the ContrastImg is adapted and the strokeWidth is constant. More...

#include <Binarization.h>

Inheritance diagram for rdf::BinarizationSuAdapted:

[legend]

Collaboration diagram for rdf::BinarizationSuAdapted:

[legend]

Public Member Functions
	BinarizationSuAdapted (const cv::Mat &img, const cv::Mat &mask=cv::Mat())
	Initializes a new instance of the BinarizationSuAdapted class. More...

virtual bool	compute () override
	Computes the binary image using the adapted method of Su. More...

virtual QString	toString () const override
	Summary of parameters as String. More...

void	setFiltering (bool medianFilter)
	If set, a median filter of size 3 is applied at the end. More...

Public Member Functions inherited from rdf::BaseBinarizationSu
	BaseBinarizationSu (const cv::Mat &img, const cv::Mat &mask=cv::Mat())
	Initializes a new instance of the BaseBinarizationSu class. More...

bool	isEmpty () const override
	Determines whether this instance is empty. More...

cv::Mat	binaryImage () const
	Returns the binary image. More...

void	setPreFiltering (bool preFilter=true, int preFilterSize=10)
	Sets the preFiltering. All blobs smaller than preFilterSize are removed if preFilter is true. More...

QSharedPointer< BaseBinarizationSuConfig >	config () const

Public Member Functions inherited from rdf::Module
	Module ()
	Default constructor Initializes a new instance of the Module class. More...

virtual QString	name () const
	Returns the module's name. More...

virtual void	setConfig (QSharedPointer< ModuleConfig > config)

QSharedPointer< ModuleConfig >	config () const

Protected Member Functions
float	setStrokeWidth (float strokeW)

virtual float	thresholdVal (float mean, float std) const

virtual float	contrastVal (const unsigned char maxVal, const unsigned char minVal) const override

virtual void	calcFilterParams (int &filterS, int &Nm) override

Protected Member Functions inherited from rdf::BaseBinarizationSu
cv::Mat	compContrastImg (const cv::Mat &srcImg, const cv::Mat &mask) const

cv::Mat	compBinContrastImg (const cv::Mat &contrastImg) const

virtual float	strokeWidth (const cv::Mat &contrastImg) const

void	computeDistHist (const cv::Mat &src, QList< int > maxDiffList, QList< float > localIntensity) const

void	computeThrImg (const cv::Mat &grayImg32F, const cv::Mat &binContrast, cv::Mat &thresholdImg, cv::Mat &thresholdContrastPxImg)

bool	checkInput () const override

Protected Member Functions inherited from rdf::Module
QString	debugName () const

Protected Attributes
cv::Mat	mContrastImg

cv::Mat	mBinContrastImg

cv::Mat	mThrImg

Protected Attributes inherited from rdf::BaseBinarizationSu
cv::Mat	mSrcImg

cv::Mat	mBwImg

cv::Mat	mMask

bool	mPreFilter = true

int	mPreFilterSize = 10

float	mStrokeW = 4

Protected Attributes inherited from rdf::Module
QSharedPointer< ModuleConfig >	mConfig

Detailed Description

The class binarize a grayvalue image. The segmentation algorithm is based on "Binarization of Historical Document Images Using Local Maximum and Minimum", Bolan Su, Shijian Lu and Chew Lim Tan, DAS 2010. In contrast to DkSegmentationSu the ContrastImg is adapted and the strokeWidth is constant.