VanGog's profile - activity

I am trying this

void copy(){
    Rect roi;           
    auto primarySegment = main_layers.begin();
    for (int c = 0; c< primaryKernelsLoad; c++)
        {
        if (heightPriority)
          {
          roi = cv::Rect(0, c, size.width, segment1Size);
          source(roi).copyTo(&(*primarySegment));
          auto nx = std::next(primarySegment, 2);
          }
        };
};

error: error C2664: 'void cv::Mat::copyTo(cv::OutputArray) const' : cannot convert parameter 1 from 'cv::Mat *' to 'cv::OutputArray'

I cannot find out what is difference between cv::OutputArray and cv::Mat and how to access the cv::OutputArray from the cv::Mat?

But you could merge the channels and then you should got RGB image which is blured, don't you? Would the results be different?

There is an interesting article "The Foundations for Scalable Multi-core Software in Intel® Threading Building Blocks" http://citeseerx.ist.psu.edu/viewdoc/... with fibonacci number example similar as is in example code within TBB package. See page 54. It's long article but contains precious examples, explaining, diagrams and benchmarks. Link to LBerger's note: http://answers.opencv.org/question/15...

What I look for is to create a loop with a listener function which will obtain a signal when some thread function have finished. I cannot imagine how it works. There should be some feeding function running in a loop in memory, which will create queue when it is needed or will pass the images from queue to the thread function. I just cannot imagine how to manage the feeding process. I will create question for this tomorrow because I just compiled OpenCV with TBB and I am completely new to this. So I first need to try your code and then I will ask.

Thanks for your answer, it's great to see example. But how can you detect that some thread has finished its job and should process next image?

@LorenaGdL: It should be possible to force range 256 based on this expression: int hrange = depth == CV_32F ? 360 : code == CV_BGR2HSV || code == CV_RGB2HSV || code == CV_BGR2HLS || code == CV_RGB2HLS ? 180 : 256; It is the 3rd parameter of function cvtColor

With Visual Studio 2010.

When I want compile OpenCV with TBB, should I add libraries of TBB to project or the settings should be prepared after cmake generated project files? When I try to compile OpenCV I got lots of messages complaining missing libraries. But I cannot change the projects. I tried to right click and add references but there is not this enabled. I can view references, not add. I cannot find the options to set compiler. I can view property manager and right click on properties but of profile Microsoft.Cpp.Win32.user but there are not compiler settings.

When I try to compile core project I got the message LINK : fatal error LNK1104: cannot open file 'tbb_debug.lib'

Update: I tried to open BUILD_ALL project and checked project settings in core project and here is possible to add the libraries to linker. I will try, but I hope I will not need to do this for every project.

I solved it. The path needs to use slashes / not \

@pklab: Now I want to recompile OpenCV with TBB and I stuck this problem. What is opencv_config_file_include_dir? Should it be the destination directory or source directory? When I started cmake GUI there is set destination directory (the last settings used). And the path is U:/OpenCV Build/3.1 win32 + contrib ... and error I get is: CMake Error at cmake/OpenCVModule.cmake:260 (foreach) - syntax error during parsing string: U:/opencv-master/modules;EXTRA;U:\opencv_contrib-master\modules ... so I dont know where does it take the EXTRA word from.

I noticed there is great difference in performance if we use filters on grayscale and colored images. This is an example code

#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/highgui/highgui.hpp"
#include <iostream>
#include <stdlib.h>
#include <stdio.h>

using namespace cv;

int main( int argc, char** argv )
{
    Mat src, gray, dst, abs_dst;
    src = imread( "../../data/lena.jpg" );
    if ( src.empty() )
        return -1;
    /// Remove noise by blurring with a Gaussian filter

    double t = (double) cv::getTickCount();
    GaussianBlur( src, dst, Size(3,3), 0, 0, BORDER_DEFAULT );
    cvtColor( dst, gray, CV_RGB2GRAY );
    t = (double) 1000 * (cv::getTickCount() - t) / cv::getTickFrequency();
    std::cout << "convertion + blur time: " << t << "ms" << std::endl;
    imshow("blured->gray", gray);
    waitKey(0);

    t = (double) cv::getTickCount();
    cvtColor( dst, gray, CV_RGB2GRAY );
    GaussianBlur( gray, gray, Size(3,3), 0, 0, BORDER_DEFAULT );
    t = (double) 1000 * (cv::getTickCount() - t) / cv::getTickFrequency();
    std::cout << "convertion + blur time: " << t << "ms" << std::endl;
    imshow("gray->blured", gray);
    waitKey(0);
    destroyWindow("blured->gray");
    destroyWindow("gray->blured");

    /// Apply Laplace function
    t = (double) cv::getTickCount();
    Laplacian( src, dst, CV_16S, 3, 1, 0, BORDER_DEFAULT );
    convertScaleAbs( dst, abs_dst );
    t = (double) 1000 * (cv::getTickCount() - t) / cv::getTickFrequency();
    std::cout << "Laplacian time: " << t << "ms" << std::endl;
    imshow( "Laplacian", abs_dst );

    waitKey(0);
    return 0;
}

So I wonder why should we perform filters on RGB images? Should we or not? I have seen various effects on grayscale images and always you convert color image to grayscale. But if it is faster to apply the filter on grayscale, could we just apply the filter on the grayscale image and then use some trick/effect that will do similar change (blur/smoothing/emboss/edges effects) to the colored images. I think like it should be many times faster than applying kernels on color images. I think like using grayscale image like some kind of "mask" or something what can change the look of the color image.

Thanks for explanation! So what image type RGB should I use exactly, not to lose it? CV_8U1, CV_8U3 will cause data lost. Do I need to use floats? So basicly I can use HSV but I need to use wider data type. But which one?

I read that OpenCV uses 0-179° of Hue. In this link the best voted answer is that docs state that OpenCV uses 0-360°. But I did not find that paragraph stating that. What is thruth? I guess the fist. But main reason why I create this question, is it save to use this namespace or can I lose information. I often use Photoshop where the range is 360° or RGB o CMYK image. So if read the image savee using iRGB ICC profile, and then read it by OpenCV and then convert it to HSL, work with it, convert it back to RGB and save it, did I lost the information? And if I measure color ranges in Photoshop and then use color ranges for HSV image (converted from RGB) do I loose information. And what if I want to work with 360° degrees, is there any way how to convert to some different data type or namespace which will contain all the data and not lose it? I am used to work with HSV colorspace because this is the most "user friendly" colorspace, it is quite simple to change lightness or hue or saturation in this colorspace so this is why I am worried about inaccuracies.

great answer, thanks

I will use float, this was for a testing purpose. I will find another way how to do it more exactly using 360 as a maximum value.

Yeah, but I need mathematical accurancy. If you use image editor you can see the image is blured in the 358-359° and that is problem. I can correct it in image editing software but it would be useless when I would use it as a build in solution. I have done recalculation for the smallest image (my question been updated). Just it is not perfect because I used floats and divisions of floats so it was a bit complicated.