Mailman 3 June 2014 - scikit-image

Re: Contribution: circles and hough transform
by François 18 Jun '14

18 Jun '14

Hi, FYI, I plan to work on the detection of ellipses quite soon. Stefan indicated to me a reference which seems fairly simple to implement. Cheers, François.

3 3

Re: Questions about tomography and scikit-image contribution
by Stéfan van der Walt 18 Jun '14

18 Jun '14

On Wed, Jun 18, 2014 at 7:07 AM, Juan Nunez-Iglesias <jni.soma(a)gmail.com> wrote: >> 2) Is there a reason (limitation) why the current radon/iradon functions >> are not more generalized? > > I don't know — I will leave this question to scikit-imagers with experience > in that module. But usually the answer to this question is: we haven't got > round to it. =) Specifically, Jostein Bø Fløystad is your man. >> 4) Where do you think the scikit-image project is in its lifecycle? > > I would say we are one or two minor releases (0.11, 0.12) from announcing > 1.0! I was going to say: "hopefully closer to the beginning than to the end" :) Regards Stéfan

1 0

Questions about tomography and scikit-image contribution
by Brian Breitsch 18 Jun '14

18 Jun '14

Hello! I am a graduate student who studies the ionosphere. This summer, I have been asked to develop a code base using python/scipy for ionospheric tomography. I have so far used the *radon* and *iradon* functions that are part of the scikit-image library. I will soon finish my preliminary research/paper-reading/algorithm-testing phase, and am starting to plan the code phase. I wonder if the code I develop may be of value to the scikit-image library. For instance, I want/need to implement a more generalized Radon transform function that allows for specifying specific angle/offset pairs, and can work with sparse matrices. Also, perhaps a function for generating projection matrices. If it it may be of value, I would obviously want to design the code from the outset to conform to the scikit testing/style/code standards and best practices (this is probably a good idea regardless). My sentiment is really similar to Emmanuelle Gouillart's: https://groups.google.com/forum/#!searchin/scikit-image/tomography/scikit-i… Although, I am surely less experienced. Here are four questions: 1) Is scikit-image the correct library for such functionality (like generalized Radon transform and projection matrix creation)? 2) Is there a reason (limitation) why the current radon/iradon functions are not more generalized? 3) If you think this functionality may be of use--any advice for development etc? 4) Where do you think the scikit-image project is in its lifecycle? Thank you so much! - Brian

2 1

Re: ANN: scikit-image 0.10.0
by Matthew Brett 14 Jun '14

14 Jun '14

Yo, On Sunday, June 1, 2014 3:35:24 AM UTC-7, Stefan van der Walt wrote: > > On Fri, May 30, 2014 at 2:26 AM, Matthew Brett <matthe...(a)gmail.com > <javascript:>> wrote: > > Congratulations? Any chance of putting up some OSX wheels on pypi? > > Sorry, folks, I've been living in a terminal for the past few days. > > This is quite a landmark: our first release that comes with a paper! > Johannes, thanks very much for handling the logistics. > > Matthew, thanks for those Wheels--can I simply upload them by hand? > Also, you pointed us to the recipes, so is the idea that we include > scripts in the project to build them automatically next time? > I noticed you only put up the Python 2.7 wheels, not the Python 3.3 and 3.4 wheels - any reason for that? Cheers, Matthew

2 3

Re: Real-time interaction forum
by Michael Aye 13 Jun '14

13 Jun '14

Hey Stéfan, would you care to share what you disliked with gitter.im? I'm asking b/c I'm also looking for a less posty, more chatty exchange forum for my projects, and since github is providing free private repos for scientists, I'm doing more and more (and more) on github, so I thought a chat client that links github projects to chat rooms is a fantastic idea. So what's wrong with it? ;) On Sunday, May 25, 2014 3:59:33 AM UTC-7, Stefan van der Walt wrote: > > After (too much) experimentation, I've come to the conclusion that > Gitter may, after all, not be ready for prime time yet. > > I suggest that we use #scikit-image on Freenode for now. I'll deploy > a Hubot for us there, and add logging ASAP. If that doesn't work, I'm > going to have to ask Stuart to help us out some more. > > Stéfan >

2 1

Re: GSoC project request for comments
by Vighnesh Birodkar 13 Jun '14

13 Jun '14

I am on it :) .

1 0

Re: GSoC project request for comments
by Stéfan van der Walt 13 Jun '14

13 Jun '14

Okay, you've convinced me. Let's implement some code with the assumption that nx is fine as a dependency and see where it takes us.

1 0

Re: GSoC project request for comments
by Juan Nunez-Iglesias 12 Jun '14

12 Jun '14

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(a)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(a)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(a)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(a)googlegroups.com. > For more options, visit https://groups.google.com/d/optout. >

1 0

Re: GSoC project request for comments
by Josh Warner 12 Jun '14

12 Jun '14

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.

2 1

Re: GSoC project request for comments
by Josh Warner 12 Jun '14

12 Jun '14

@Ronnie If I’m reading the docs correctly, the only *algorithmic* requirements for NetworkX which matter for programmatic use are Python >= 2.6, NumPy, and SciPy, which we also require. The optional packages only provide visualization capabilities or the ability to read new file formats, which would not be necessary for our purposes. http://networkx.github.io/documentation/latest/install.html

2 1