Moster's profile - activity

And it goes on...

void doEM1D(Mat& _im, Mat& probs,int num_models = 2,bool useRanges = true,Mat& lables = Mat()) {
    //Mat im; cvtColor(_im,im,CV_RGB2HSV);
    Mat im = _im;

    vector<vector<CvEM>> model(num_models);
    for(int i=0;i<num_models;i++) {
            model[i] = vector<CvEM>(3);
    }
    CvEMParams ps(1);

error C2065: 'CvEM' : undeclared identifier error C2923: 'std :: vector': 'CVEM' is not a valid type template argument for parameter '_Ty' error C2065: 'CvEMParams' : undeclared identifier

Did you try to simply print out the type of the src Mat?

The type you can find here: https://github.com/Itseez/opencv/blob/master/modules/core/include/opencv2/core/cvdef.h (line 258). So 0 means 8 bit 1 channel. To get the higher channel equivalents, you just use this formula: type_number + (channels - 1) * 8.

So, for example 8 bit 3 channel would be simply 16.

http://docs.opencv.org/modules/core/doc/operations_on_arrays.html#sum This of course only works if you want the sum. Since you also calculated the square in your example, you need to have a loop. The java version seems to be this: http://docs.opencv.org/java/2.4.2/org/opencv/core/Core.html#sumElems%28org.opencv.core.Mat%29

http://stackoverflow.com/questions/12394364/opencv-for-android-access-elements-of-mat

http://docs.opencv.org/modules/features2d/doc/feature_detection_and_description.html The first one is usually the default. You also find it in the source at line 360 https://github.com/Itseez/opencv/blob/master/modules/features2d/src/fast.cpp

hi, how can i add each pixel of a matrix in a loop in android

below is my c++ code :

for(int x=0;x<=4;x++)
{
    for(int y=0;y<=4;y++)
    {
    sum1=sum1+ block.at<uchar>(x, y);  // Addition of all elements
    sum2=sum2+block.at<uchar>(x,y)*block.at<uchar>(x,y); //Addition of all element square                   
    }
}

where block is a 5x5 mat

It tests for 9 contiguous pixels that are brighter or darker. That is the standard.

And obviously also the standard deviation as the name says. "Calculates a mean and standard deviation of array elements."

I today you want I'm your Connie 100 back you UK I'm thinking and anything looking 15 and me we have a whole are hey here like of he have yeah okay it go all the way always hitting up you reveal everything K yeah your lives on %uh job slow I hate me %uh all-male down yeah it no gay up or me Hill it hard 3 hey last year and one Helms’s yes and I'm getting old will be doing second affecting heating up okay well hey I him shop class day him I me only our and crunching caved gauge it go I'll any yo

Street from the back the ballad never going to leave that ever gain didn't leave on Telegraph actual you know when I was Ryan's any they were only ever yes I did many basketball every day press in law I’m great because she P much every time ago that we all eat something called for pretty much missed saw Dyes so you can get a dice a call on rest deal of then can do that show construction but still the whip other day a salary is when doesn't understand what he's tall his fruit we skip a meal East meal to die with me he was smashed and a hot more gaining seller witnessing say all possible I'll tell you he nodded for to go to steam for could be against us chart up see what's the I want all a little time he would I will it all and that's why popular is what we is away this is what we festival we actually ate got home from that were you know it nice song Don has excellen

Edit: Links removed

It does work forth youth in fact it is the most to set something that parents need to be teaching their children they can't be left out a parenting and it's almost easier for the these kids to get it then it is the adults I like watching them because they'll they all respond more quickly now I wanton watch this video from Zion kids they'll give you hope if you have child that struggling here's hope for your child Jesus Christmas humble teach loyalty thankfulness truthfulness on a courage and devotion teach them to be unified with others who are here for God's kingdom and you have the key McGowan in their hearts .

No error message, no code. How should we help?

16 bit signed is also supported. I dont know the exact reason, but Im sure its related to the color spaces. Color spaces are usually represented in one of those 3, so 64 bit floats are not meant to be supported for image processing.

No, sorry. I never messed with videoediting on android. Isnt there an android API for that?

No, you cannot. Opencv does not support reading/writing videos on android (yet).

Its not explicitly mentioned, but the return value should give you a clue if the set was successful. Its a simple boolean.

Usually python is indeed a lot slower, but in opencv python just wraps around the c/c++ code of opencv. So running the opencv operations shouldnt be slower at all. You dont know exactly what the guy in the video did. Maybe he also used the CUDA surf implementation of opencv to reach less computation time. Anyways, you can of course try your simple example in c++. It shouldnt be that hard, since the tutorials almost contain everything you need already.

FAST and FASTX are simply functions and not to be used as constructors. So just call them with the given 4/5 arguments.

I've got two ways to capture a camera. 1.using the functions in the opencv

 #include < opencv\cv.h
 #include < opencv\highgui.h 
 #include < iostream 
 using namespace std;
 using namespace cv;
 int main()
 {
     char c;
     CvCapture* capture = cvCreateCameraCapture(0);   //0 is the external camera ID，1 is the internal camera ID of my laptop.
     IplImage* src;
     for (;;)
    {
        src = cvQueryFrame(capture);         
        cvShowImage("Input", src);
        c = waitKey(10);
        if (c == 27) break;
    } 
   cvReleaseCapture(&capture);
   return 0;
 }
2.using the functions in the ***opencv2***
 #include < opencv2\core\core.hpp
 #include < opencv2\highgui\highgui.hpp 
 #include < iostream 
 using namespace cv;
 using namespace std;
 int main()
 {
     char c;
     VideoCapture inputVideo(0);    //0为外部摄像头的ID，1为笔记本内置摄像头的ID
     Mat src;
     for(;;)
     {
        inputVideo >> src;
        imshow("input",src);
        c = waitKey(10);
        if (c == 27) break;
     }
   return 0;
 }

By using the method 1, the program can read both the external and the internal camera's data;but the program can only read the internal camera's data with method 2. I don't know why method 1 and method 2 so different.
What should I do if I want to get the data of the external camera with method 2?
// my operation system is Window8.1, the version of my OPENCV is 2.4.8, the version of my Visual Studio is 2013. I need your help.

Your threshold for the "good features extraction" is only min_dist, but should be something like 2 or 3 times min_dist. Its kind of obvious that you only get a few matches if the distance should be equal to min_dist.

for(int j = 0; j < desc_left.rows(); j++){
    if(matchesList.get(j).distance <= 2*min_dist){
        good_matches.addLast(matchesList.get(j));
    }
}

But matlab is not a classical programming language to me. Its rather a scripting language that is also open to non-programmers, most likely a reason to start with 1 :)

Thanks, good to know. I havent touched fortran yet.

-1 usually means that this feature is simply not supported. So you are right.

Do you necessarily need to display all frames?

I dont know how long you have been programming, but since when do indices start with 1? I personally dont know any modern language that does that. So of course your matx33 object starts with (0,0) and ends with (2,2).

Matx33f woah;
woah(0,0) = 0.0;
woah(0,1) = 0.0;
woah(0,2) = 0.0;
woah(1,0) = 0.0;
woah(1,1) = 0.0;
woah(1,2) = 0.0;
woah(2,0) = 0.0;
woah(2,1) = 10.0;
woah(2,2) = 10.0;

cout << woah << endl;

Edit: Btw, one interesting thing to know. Only the Mat object has a range check as far as I know. If you use Matx or Mat_< type > you wont notice if you accidently access some random memory region.

If you use cout on uchar, you will of course get "random" signs, since it handles it as chars (I think ascii). So cast them to int with (int) and you should get numbers.

Could you add the exact error message?

Check out the sample applications in the android opencv SDK. For example, the face detection and mixed processing sample show how to do that.

Just search for Suzuki, S. and Abe, K., Topological Structural Analysis of Digitized Binary Images by Border Following. CVGIP 30 1, pp 32-46.

Necessary papers are usually linked in the documentation(also here): http://docs.opencv.org/modules/imgproc/doc/structural_analysis_and_shape_descriptors.html?highlight=findcontours#findcontours