形状距离和匹配

详细描述

类
类	cv::AffineTransformer
	OpenCV 仿射变换算法的包装类。: 更多...
类	cv::ChiHistogramCostExtractor
	基于 Chi 的成本提取。: 更多...
类	cv::EMDHistogramCostExtractor
	基于 EMD 的成本提取。: 更多...
类	cv::EMDL1HistogramCostExtractor
	基于 EMD-L1 的成本提取。: 更多...
类	cv::HausdorffDistanceExtractor
	由轮廓定义的形状之间的简单豪斯多夫距离度量。 更多...
类	cv::HistogramCostExtractor
	直方图成本算法的抽象基类。 更多...
类	cv::NormHistogramCostExtractor
	基于范数的成本提取。: 更多...
类	cv::ShapeContextDistanceExtractor
	形状上下文描述符和匹配算法的实现。 更多...
类	cv::ShapeDistanceExtractor
	形状距离算法的抽象基类。 更多...
类	cv::ShapeTransformer
	形状变换算法的抽象基类。 更多...
类	cv::ThinPlateSplineShapeTransformer
	变换的定义。 更多...

函数
Ptr< AffineTransformer >	cv::createAffineTransformer (bool fullAffine)
Ptr< HistogramCostExtractor >	cv::createChiHistogramCostExtractor (int nDummies=25, float defaultCost=0.2f)
Ptr< HistogramCostExtractor >	cv::createEMDHistogramCostExtractor (int flag=DIST_L2, int nDummies=25, float defaultCost=0.2f)
Ptr< HistogramCostExtractor >	cv::createEMDL1HistogramCostExtractor (int nDummies=25, float defaultCost=0.2f)
Ptr< HausdorffDistanceExtractor >	cv::createHausdorffDistanceExtractor (int distanceFlag=cv::NORM_L2, float rankProp=0.6f)
Ptr< HistogramCostExtractor >	cv::createNormHistogramCostExtractor (int flag=DIST_L2, int nDummies=25, float defaultCost=0.2f)
Ptr< ShapeContextDistanceExtractor >	cv::createShapeContextDistanceExtractor (int nAngularBins=12, int nRadialBins=4, float innerRadius=0.2f, float outerRadius=2, int iterations=3, const Ptr< HistogramCostExtractor > &comparer=createChiHistogramCostExtractor(), const Ptr< ShapeTransformer > &transformer=createThinPlateSplineShapeTransformer())
Ptr< ThinPlateSplineShapeTransformer >	cv::createThinPlateSplineShapeTransformer (double regularizationParameter=0)
float	cv::EMDL1 (InputArray signature1, InputArray signature2)
	根据论文“EMD-L1：一种比较基于直方图的描述符的有效且稳健的算法”，由 Haibin Ling 和 Kazunori Okuda 撰写；以及“地球移动距离是 Mallows 距离：来自统计学的一些见解”，由 Elizaveta Levina 和 Peter Bickel 撰写，计算两个加权点配置之间的“最小工作”距离。

函数文档

createAffineTransformer()

Ptr< AffineTransformer > cv::createAffineTransformer

(

bool

fullAffine

)

Python
	cv.createAffineTransformer(	fullAffine	) ->	retval

#include <opencv2/shape/shape_transformer.hpp>

完整的构造函数

createChiHistogramCostExtractor()

Ptr< HistogramCostExtractor > cv::createChiHistogramCostExtractor	(	int	nDummies = 25,
		float	defaultCost = 0.2f
	)

Python
	cv.createChiHistogramCostExtractor(	[, nDummies[, defaultCost]]	) ->	retval

#include <opencv2/shape/hist_cost.hpp>

createEMDHistogramCostExtractor()

Ptr< HistogramCostExtractor > cv::createEMDHistogramCostExtractor	(	int	flag = DIST_L2,
		int	nDummies = 25,
		float	defaultCost = 0.2f
	)

Python
	cv.createEMDHistogramCostExtractor(	[, flag[, nDummies[, defaultCost]]]	) ->	retval

#include <opencv2/shape/hist_cost.hpp>

createEMDL1HistogramCostExtractor()

Ptr< HistogramCostExtractor > cv::createEMDL1HistogramCostExtractor	(	int	nDummies = 25,
		float	defaultCost = 0.2f
	)

Python
	cv.createEMDL1HistogramCostExtractor(	[, nDummies[, defaultCost]]	) ->	retval

#include <opencv2/shape/hist_cost.hpp>

createHausdorffDistanceExtractor()

Ptr< HausdorffDistanceExtractor > cv::createHausdorffDistanceExtractor	(	int	distanceFlag = cv::NORM_L2,
		float	rankProp = 0.6f
	)

Python
	cv.createHausdorffDistanceExtractor(	[, distanceFlag[, rankProp]]	) ->	retval

#include <opencv2/shape/shape_distance.hpp>

createNormHistogramCostExtractor()

Ptr< HistogramCostExtractor > cv::createNormHistogramCostExtractor	(	int	flag = DIST_L2,
		int	nDummies = 25,
		float	defaultCost = 0.2f
	)

Python
	cv.createNormHistogramCostExtractor(	[, flag[, nDummies[, defaultCost]]]	) ->	retval

#include <opencv2/shape/hist_cost.hpp>

createShapeContextDistanceExtractor()

Ptr< ShapeContextDistanceExtractor > cv::createShapeContextDistanceExtractor	(	int	nAngularBins = 12,
		int	nRadialBins = 4,
		float	innerRadius = 0.2f,
		float	outerRadius = 2,
		int	iterations = 3,
		const Ptr< HistogramCostExtractor > &	comparer = createChiHistogramCostExtractor(),
		const Ptr< ShapeTransformer > &	transformer = createThinPlateSplineShapeTransformer()
	)

Python
	cv.createShapeContextDistanceExtractor(	[, nAngularBins[, nRadialBins[, innerRadius[, outerRadius[, iterations[, comparer[, transformer]]]]]]]	) ->	retval

#include <opencv2/shape/shape_distance.hpp>

示例

modules/shape/samples/shape_example.cpp.

createThinPlateSplineShapeTransformer()

Ptr< ThinPlateSplineShapeTransformer > cv::createThinPlateSplineShapeTransformer

(

double

regularizationParameter = 0

)

Python
	cv.createThinPlateSplineShapeTransformer(	[, regularizationParameter]	) ->	retval

#include <opencv2/shape/shape_transformer.hpp>

完整的构造函数

EMDL1()

float cv::EMDL1	(	InputArray	signature1,
		InputArray	signature2
	)

#include <opencv2/shape/emdL1.hpp>

根据论文“EMD-L1：一种比较基于直方图的描述符的有效且稳健的算法”，由 Haibin Ling 和 Kazunori Okuda 撰写；以及“地球移动距离是 Mallows 距离：来自统计学的一些见解”，由 Elizaveta Levina 和 Peter Bickel 撰写，计算两个加权点配置之间的“最小工作”距离。

参数

signature1	第一个签名，一个单列浮点矩阵。 每一行是直方图在每个 bin 中的值。
signature2	与 signature1 格式和大小相同的第二个签名。