So far we have developed the algorithm implementation for K-means clustering, with particular application to images.

The delopment is in three branches:

• singlethread, is a not parallellized development.
• cputhread, is a CPU parallellized development.
• cudabranch, is a GPU parallellized development.

Go to the cpp directory, and use the "make" command for compilation.

$ cd cpp/
$ make

You should see some like this:

g++ -O2 -Wall -Wextra -fPIC -c -o matrix.o matrix.cpp
g++ -O2 -Wall -Wextra -fPIC -c -o kmeans.o kmeans.cpp
g++ -shared -Wl,-soname,libutpann.so.1 -O2 -Wall -Wextra -fPIC -o libutpann.so matrix.o kmeans.o -lc

in this moment you have a library for K-means algorithm compilated, but is not installed yet, for this use.

# make install

Now is moment to see the K-means in action, so go to the singlethread branch and in the path cpp/test/colkm for test it with images. first make sure you have installed the requirements to process the images. You can install them with the following line.

# aptitude install libcvaux-dev libhighgui-dev libcv-dev

Developed by insilico.