benzun's profile - activity

void mouseEvent(int evt, int x, int y, int flags, void* param)
{
    IplImage* rgb = (IplImage*) param;
    if(evt==CV_EVENT_MOUSEMOVE)
    {
          char* data = rgb->imageData + rgb->widthStep * y + x * 3;
          printf("Current Position: x= %d y= %d B=%d G=%d R=%d\n",x,y,data[0],data[1],data[2]);
    }
}


void main()
{
    IplImage *image;
.... load image....
     cvSetMouseCallback(name, my_mouse_callback, (void*) image);
}

you can type cast IplImage as void and pass it to the function, That way you dont need to have it as global

String face_cascade_name = "haarcascade_frontalface_alt.xml"; String eyes_cascade_name = "haarcascade_eye_tree_eyeglasses.xml";

you need to replace these strings with the actual location of the cascades they will mostly be in the location where you installed opencv. Or copy them to your directory. If you dont point to the correct address for the cascade it wont work!

You need to pass your image as a parameter to the callback and read it from the image. Please read the tutorials on docs.opencv where a clear explanation about accessing pixel values is described!

Thanks for verifying it berak. I will mostly upgrade to newer version but now I will use my own function.

you should not spam the answers.opencv with your question!

When you compile opencv you will get opencv_createsamples and opencv_traincascade as a part of the application. This are binaries automatically created. If you installed opencv in /usr/bin they should be available in your terminal.

So far this is the most detailed and the best tutorial I have found on it http://note.sonots.com/SciSoftware/haartraining.html

What do you mean by a strange image? can you please post it?

When i try to use the norm function for the classes of the template Point_ and Point3_ I get an error that says that there is no known conversion from Point_ to InputArray. Is this a bug? Please tell me how to resolve this. I can write my own norm function but that does not solve the problem.

Also norm seems to do ok with Point but not Point2f.

I am using opencv 2.3.1 do i need to upgrade?

If you dont want to build tbb from source and are using linux you can pull the library using apt-get. libtbb2 and libtbb-dev are the libraries you need to install

@Adi usually the first camera that you plug in gets assigned as 0 and every camera you add is ennumerated+1 from there in ;inux

Even i get the same error but my camera still loads

I think this error occurs because the v4l library is not able to access all the properties of your camera

Check if there is too much noise in your image. Also check if you setting thresholds for canny properly.

Hi.

Does anyone where i can get a sample model and intrinsic parameters so I can test the example.

Thanks, Benzun