日期: 2019 年 10 月 7 日

暑假时候复现的一篇计算机视觉经典抠图算法grabcut的论文，“GrabCut” — Interactive Foreground Extraction using Iterated Graph Cuts

代码存一下，虽然我转行了不搞cv了hhh

myGrabCut.h

#pragma once

#include<graph.cpp>
#include<opencv2/opencv.hpp>
#include <opencv2/core.hpp>
#include <opencv2/core/types.hpp>
#include<string>
typedef Graph<double, double, double> GraphType;

using namespace cv;


class GMM {
public:
	Mat model;
	static const int compo_k = 5;
	double coefs[compo_k];
	double mean[compo_k][3];
	double covs[compo_k][3][3];
	double inv_covs[compo_k][3][3];
	double det_covs[compo_k];
	double sums[compo_k][3];
	double prods[compo_k][3][3];
	int sample_num[compo_k];
	int tot_sample_num;

	GMM();
	void calc_inv_det(int ci);
	double prob_GMM(const Vec3d color) const;
	double prob_Gi(int ci, const Vec3d color) const;
	int sle_compo(const Vec3d color) const;
	void init_param();
	void add_pixel(int ci, const Vec3d color);
	void esti_param();

};


static double calcBeta(const Mat& img);
	
/*
  Calculate weights of noterminal vertices of graph.
  beta and gamma - parameters of GrabCut algorithm.
 */
static void calcNWeights(const Mat& img, Mat& leftW, Mat& upleftW, Mat& upW,
	Mat& uprightW, double beta, double gamma);

/*
  Check size, type and element values of mask matrix.
 */
static void checkMask(const Mat& img, const Mat& mask);

/*
  Initialize mask using rectangular.
*/
static void initMaskWithRect(Mat& mask, Size imgSize, Rect rect);

/*
  Initialize GMM background and foreground models using kmeans algorithm.
*/
static void initGMMs(const Mat& img, const Mat& mask, GMM& bgdGMM, GMM& fgdGMM);

/*
  Assign GMMs components for each pixel.
*/
static void assignGMMsComponents(const Mat& img, const Mat& mask, const GMM& bgdGMM,
	const GMM& fgdGMM, Mat& compIdxs);

//论文中：迭代最小化算法step 2：从每个高斯模型的像素样本集中学习每个高斯模型的参数
/*
  Learn GMMs parameters.
*/
static void learnGMMs(const Mat& img, const Mat& mask, const Mat& compIdxs, GMM& bgdGMM, GMM& fgdGMM);

/*
  Construct GCGraph
*/
static void constructGCGraph(const Mat& img, const Mat& mask, const GMM& bgdGMM, const GMM& fgdGMM, double lambda,
	const Mat& leftW, const Mat& upleftW, const Mat& upW, const Mat& uprightW,
	GraphType*& graph);

/*
  Estimate segmentation using MaxFlow algorithm
*/
static void estimateSegmentation(GraphType* graph, Mat& mask, std::vector<double>& result);

void mygrabCut(InputArray _img, InputOutputArray _mask, Rect rect,
	InputOutputArray _bgdModel, InputOutputArray _fgdModel,
	int iterCount, int mode);//,std::vector<double>& result);

主程序代码，其中最大流最小割直接用了maxflow-v3.01的库，其他调用了一些OpenCv的基本接口我不会说我就是照着OpenCv的库函数写的
继续阅读“[CV][经典算法] Grab Cut C++实现”