标签:views word string board 不同的 scale 实现 ros 算法
https://blog.csdn.net/cai13160674275/article/details/72991049?locationNum=7&fps=1
上文讲了几种简单的方法,显著性检测就是把一幅图像中最吸引人注意的部分提取出来。
我用opencv重写了LC,AC,FT三种算法,代码和效果如下:
1.,后面的方法其实大概都是基于这个实现的,代码样子差不多
LC思路就是利用对某个像素点累加其与全幅像素的距离(欧式距离),然后归一化到0-255,由于是rgb信息,于是用直方图优化,提前计算好每个颜色与其他的距离和
- void SalientRegionDetectionBasedonLC(Mat &src){
- int HistGram[256]={0};
- int row=src.rows,col=src.cols;
- int gray[row][col];
-
- int val;
- Mat Sal=Mat::zeros(src.size(),CV_8UC1 );
- Point3_<uchar>* p;
- for (int i=0;i<row;i++){
- for (int j=0;j<col;j++){
- p=src.ptr<Point3_<uchar> > (i,j);
- val=(p->x + (p->y) *2 + p->z)/4;
- HistGram[val]++;
- gray[i][j]=val;
- }
- }
-
- int Dist[256];
- int Y,X;
- int max_gray=0;
- int min_gray=1<<28;
- for (Y = 0; Y < 256; Y++)
- {
- val = 0;
- for (X = 0; X < 256; X++)
- val += abs(Y - X) * HistGram[X];
- Dist[Y] = val;
- max_gray=max(max_gray,val);
- min_gray=min(min_gray,val);
- }
-
-
- for (Y = 0; Y < row; Y++)
- {
- for (X = 0; X < col; X++)
- {
- Sal.at<uchar>(Y,X) = (Dist[gray[Y][X]] - min_gray)*255/(max_gray - min_gray);
-
-
- }
- }
- imshow("sal",Sal);
- waitKey(0);
-
- }
效果图
2.AC算法也挺有意思,利用了类似图像金字塔算法,在不同纬度(具体实现是用大小不同的均值滤波器过滤图像)与标准图像做差并累加,然后归一化,实现如下:
- void SalientRegionDetectionBasedonAC(Mat &src,int MinR2, int MaxR2,int Scale){
- Mat Lab;
- cvtColor(src, Lab, CV_BGR2Lab);
-
- int row=src.rows,col=src.cols;
- int Sal_org[row][col];
- memset(Sal_org,0,sizeof(Sal_org));
-
- Mat Sal=Mat::zeros(src.size(),CV_8UC1 );
-
- Point3_<uchar>* p;
- Point3_<uchar>* p1;
- int val;
- Mat filter;
-
- int max_v=0;
- int min_v=1<<28;
- for (int k=0;k<Scale;k++){
- int len=(MaxR2 - MinR2) * k / (Scale - 1) + MinR2;
- blur(Lab, filter, Size(len,len ));
- for (int i=0;i<row;i++){
- for (int j=0;j<col;j++){
- p=Lab.ptr<Point3_<uchar> > (i,j);
- p1=filter.ptr<Point3_<uchar> > (i,j);
-
-
- val=sqrt( (p->x - p1->x)*(p->x - p1->x)+ (p->y - p1->y)*(p->y-p1->y) + (p->z - p1->z)*(p->z - p1->z) );
- Sal_org[i][j]+=val;
- if(k==Scale-1){
- max_v=max(max_v,Sal_org[i][j]);
- min_v=min(min_v,Sal_org[i][j]);
- }
- }
- }
- }
-
- cout<<max_v<<" "<<min_v<<endl;
- int X,Y;
- for (Y = 0; Y < row; Y++)
- {
- for (X = 0; X < col; X++)
- {
- Sal.at<uchar>(Y,X) = (Sal_org[Y][X] - min_v)*255/(max_v - min_v);
-
- }
- }
- imshow("sal",Sal);
- waitKey(0);
- }
SalientRegionDetectionBasedonAC(test,test.rows/8,test.rows/2,3);
3.FT算法
lab空间的均值减去当前像素值
- void SalientRegionDetectionBasedonFT(Mat &src){
- Mat Lab;
- cvtColor(src, Lab, CV_BGR2Lab);
-
- int row=src.rows,col=src.cols;
-
- int Sal_org[row][col];
- memset(Sal_org,0,sizeof(Sal_org));
-
- Point3_<uchar>* p;
-
- int MeanL=0,Meana=0,Meanb=0;
- for (int i=0;i<row;i++){
- for (int j=0;j<col;j++){
- p=Lab.ptr<Point3_<uchar> > (i,j);
- MeanL+=p->x;
- Meana+=p->y;
- Meanb+=p->z;
- }
- }
- MeanL/=(row*col);
- Meana/=(row*col);
- Meanb/=(row*col);
-
- GaussianBlur(Lab,Lab,Size(3,3),0,0);
-
- Mat Sal=Mat::zeros(src.size(),CV_8UC1 );
-
- int val;
-
- int max_v=0;
- int min_v=1<<28;
-
- for (int i=0;i<row;i++){
- for (int j=0;j<col;j++){
- p=Lab.ptr<Point3_<uchar> > (i,j);
- val=sqrt( (MeanL - p->x)*(MeanL - p->x)+ (p->y - Meana)*(p->y-Meana) + (p->z - Meanb)*(p->z - Meanb) );
- Sal_org[i][j]=val;
- max_v=max(max_v,val);
- min_v=min(min_v,val);
- }
- }
-
- cout<<max_v<<" "<<min_v<<endl;
- int X,Y;
- for (Y = 0; Y < row; Y++)
- {
- for (X = 0; X < col; X++)
- {
- Sal.at<uchar>(Y,X) = (Sal_org[Y][X] - min_v)*255/(max_v - min_v);
-
- }
- }
- imshow("sal",Sal);
- waitKey(0);
- }
简单的图像显著性区域特征提取方法-----opencv实现LC,AC,FT
标签:views word string board 不同的 scale 实现 ros 算法
原文地址:https://www.cnblogs.com/jukan/p/9209236.html