For the majority of users, nx will already be installed as part of their scientific python distribution (Anaconda, Canopy, Python-xy, etc). For the rest, nx's install size is small enough that I don't think it's an issue. Stéfan's question is more pertinent. What does nx provide that we need or would find useful? - max flow / min cut useful for some segmentation algorithms - minimum spanning trees (maximin affinity learning) - biconnected components (useful to remove segments embedded in other segments) - shortest paths - community detection There's probably other graph-theoretical nicenesses that are useful for segmentation but I haven't considered yet. scipy.sparse.csgraph provides a fraction of these. And our own custom class provides none of these. Juan. On Fri, Jun 13, 2014 at 9:24 AM, Ronnie Ghose <ronnie.ghose@gmail.com> wrote:

sorry, i mean required dependency in the sense of when i say install blah, well after everything, i have this much memory now occupied. so it would be scikit + networkx size now... :/

On Thu, Jun 12, 2014 at 7:19 PM, Josh Warner <silvertrumpet999@gmail.com> wrote:

...

To be clear: when I say required dependency I mean requiring it as an install and run time dependency on the user's system, like we require NumPy and SciPy. Not adding it to our repo. This would not increase the size of scikit-image by more than a few bytes.

-- You received this message because you are subscribed to the Google Groups "scikit-image" group. To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image+unsubscribe@googlegroups.com. For more options, visit https://groups.google.com/d/optout.

-- You received this message because you are subscribed to the Google Groups "scikit-image" group. To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image+unsubscribe@googlegroups.com. For more options, visit https://groups.google.com/d/optout.