[1] M. van den Bergh, X. Boix, G. Roig, B. de Capitani, L. van Gool.
    SEEDS: Superpixels extracted via energy-driven sampling.
    Proceedings of the European Conference on Computer Vision, pages 13–26, 2012.

[2] D. Stutz, A. Hermans, B. Leibe.
    Superpixel Segmentation using Depth Information.
    Bachelor thesis, RWTH Aachen University, Aachen, Germany, 2014.

[3] P. Arbeláez, M. Maire, C. Fowlkes, J. Malik.
    Contour detection and hierarchical image segmentation.
    Transactions on Pattern Analysis and Machine Intelligence, 33(5):898–916, 2011.
[4] N. Silberman, D. Hoiem, P. Kohli, R. Fergus.
    Indoor segmentation and support inference from RGBD images.
    Proceedings of the European Conference on Computer Vision, pages 746–760, 2012.

# Prerequisites assuming OpenCV is already installed.
sudo apt-get install build-essential cmake libboost-dev-all
# Clone the repository:
$ git clone https://github.com/davidstutz/seeds-revised.git
# Go to the build subfolder to generate the CMake files:
$ cd seeds-revised/build
$ cmake ..
# Compile the library and corresponding Command Line Interface:
$ make

$ ../bin/cli --help
Allowed options:
    --help                          produce help message
    --input arg                     the folder to process, may contain several 
                              images
    --bins arg (=5)                 number of bins used for color histograms
    --neighborhood arg (=1)         neighborhood size used for smoothing prior
    --confidence arg (=0.100000001) minimum confidence used for block update
    --iterations arg (=2)           iterations at each level
    --spatial-weight arg (=0.25)    spatial weight
    --superpixels arg (=400)        desired number of supüerpixels
    --verbose                       show additional information while processing
    --csv                           save segmentation as CSV file
    --contour                       save contour image of segmentation
    --labels                        save label image of segmentation
    --mean                          save mean colored image of segmentation
    --output arg (=output)          specify the output directory (default is 
                              ./output)

#include <opencv2/opencv.hpp>
#include "SeedsRevised.h"
#include "Tools.h"

// ...

cv::Mat image = cv::imread(filepath);

// Number of desired superpixels.
int superpixels = 400;

// Number of bins for color histograms (per channel).
int numberOfBins = 5;

// Size of neighborhood used for smoothing term, see [1] or [2].
// 1 will be sufficient, >1 will slow down the algorithm.
int neighborhoodSize = 1;

// Minimum confidence, that is minimum difference of histogram intersection
// needed for block updates: 0.1 is the default value.
float minimumConfidene = 0.1;

// The weighting of spatial smoothing for mean pixel updates - the euclidean
// distance between pixel coordinates and mean superpixel coordinates is used
// and weighted according to:
//  (1 - spatialWeight)*colorDifference + spatialWeight*spatialDifference
// The higher spatialWeight, the more compact superpixels are generated.
float spatialWeight = 0.25;

// Instantiate a new object for the given image.
SEEDSRevisedMeanPixels seeds(image, superpixels, numberOfBins, neighborhoodSize, minimumConfidence, spatialWeight);

// Initializes histograms and labels.
seeds.initialize();
// Runs a given number of block updates and pixel updates.
seeds.iterate(iterations);

// Save a contour image to the following location:
std::string storeContours = "./contours.png";

// bgr color for contours:
int bgr[] = {0, 0, 204};

// seeds.getLabels() returns a two-dimensional array containing the computed
// superpixel labels.
cv::Mat contourImage = Draw::contourImage(seeds.getLabels(), image, bgr);
cv::imwrite(store, contourImage);

CV_BGR2GRAY
CV_BGR2Lab