basarkir's profile - activity

I have used ORB for keypoints and descriptors but I want to try new algorithms for detection.ORB is good but sometimes it getting worse for detection.Now I am testing freak.I use fast for keypoint extracting.But it is slow while matching (BruteForceMatcher<hamming>).Do you have any ideas about a good matcher for fast and freak ?

I use "perspectiveTransform(inPts, outPts, homography);" and it works well.Thank you.

I have two points.Train points are (x,y) and test points are (x',y').I want to muliply xhomography and yhomography.For a result ,I want to get x' and y'.

I want to multiply two points and matrix.I have trainPoints which is trained before the program.And I have testPoints which is occured while program is processing.And also I have homography which is Mat object (calculated with findHomography method).I want to check is it true?I want to calculate homography error.How can i multiply two points and homography?

Thank you so much Micheal for your answer. I found solution. Here is my working code:

void roi(Mat& img,int x_point,int y_point,int width ,int height)
{
    Rect roi = Rect(x_point,y_point,width,height);
    Mat roiImg;
    roiImg = img(roi);
    equalizeHist(roiImg,roiImg);
    roiImg.copyTo(img(Rect(x_point, y_point, width, height)));
}

I use histogram for illumination images but I wanted to use it only on the center of the image because of time limitations. My roi function's execute time is 0.2ms. Using the equalizeHist function's execute time is 1.0ms. In addition to this, when I use equalizeHist, it sometimes causes noises. You gave me some useful information about images. I have always used cvtColor, but now I am going to use imread method with (1,0,-1).

I want to use histogram only the center part of my image.I use histogram for illumination images but I wanted to use it only on the center of the image because of time limitations. In addition to this, when I use equalizeHist, it sometimes causes noises.Here is my code;

Mat img1=imread("imageLight100.bmp");
Mat img2=img1.clone();
Rect roi = Rect(100,100,200,200);
Mat roiImg;

roiImg = img1(roi);
cvtColor(roiImg, roiImg, CV_RGB2GRAY);
equalizeHist(roiImg,roiImg);
imshow("roi",roiImg);

cvtColor(img2, img2, CV_RGB2GRAY);
img2.copyTo(roiImg);
imshow("roiImage",img2);
waitKey();

I want to know ,why it is too slow? And how can I speed up it?

tracker Prepare time= 3.328474 tracker applyOF time= 0.010696 Detection Prepare time= 2.133963 Freak num of keypoints: 969 FREAK Test Time= 94.338517 Freak num of descriptors: 882 matching time= 5.424213 Number of matches object 1 : 61 Tracker homography time= 146.425174 Number of matches object 2 : 141 Tracker homography time= 9.664674 Detection PostProcess time= 9.719263 execute time= 262.028259

freak test time is Freak's compute method matching time is my matching method homography time is computeHomography method

i tried it on ubuntu ,and i think if it is slow on ubuntu there is no sense to use it on android =) i work about augmented reality ,when it find nearly 150 keypoints ,tracking'll take 140 ms.

which way did you decide to go on your project ? did u return to ORB?

I'm using ORB on my project and I have decided to change my extractor to freak with opencv 2.4.2. Now I'm working on Freak but it works slower than ORB. I don't understand this, because the paper about freak says: "it's very fast (comparable to ORB)"

I used Fast and Surf for keypoints extracting, both of them are slow. Do you have any idea about it?