Using k-means cluster algorithm to computing the dominant colors of given image.

• Copy all .cpp and .h files except main.cpp to your project.

• include mashiro.h

• Init class mashiro with an instance of cv::Mat

    Mat image = imread("/PATH/TO/AN/IMAGE");
    mashiro mashiro(image);
  

• Call color function

    mashiro.color(3, [](cv::Mat& image, Cluster colors){
        for_each(colors.cbegin(), colors.cend(), [](const MashiroColor& color){
            cout<<"("
                <<color[0]<<", "
                <<color[1]<<", "
                <<color[2]<<")"
            <<endl;
        });
    });
  

  The first parameter is the k in K-means cluster algorithm

  The second parameter is a callback function, which gives the reference of the input image and clustered colors.

Just compile and install it with

$ make && make install
$ mashiro
Usage:
	-i [image file] -c [number of color to cluster]
	-h Print this help

Notice,

In Makefile, LIB and INCLUDE might need to be changed to give correct path to OpenCV.

mashiro -i cover.jpg -c 3

If everything is fine, you'll see

My blog post