This notebook introduces some basic techniques for image segmentation, such as the use of k-means -a simple yet extremely popular clustering algorithm (also considering its iterative version)- and superpixels algorithm such as SLIC, for applying the Normalized Cut to the obtained Region Adjacency Graph.
You can download the sample dataset here (from microsoft website).
For more information on k-means implementation, check out the scikit website.
For further insight into SLIC (superpixeling algorithm), check out the scikit-image implementation, as well as the original paper.
For additional details about the Normalized Cut algorithm, refer once again to the original paper.