Re: Relation with computer vision and machine learning
Hi Wei.
Thanks for your reply! I see your great work in scikits-learn and your comments are quite useful.
Thanks :)
Many will overlap, I think we can maintain a list. As a quick example in the following: There is no discussion that machine learning methods are helpful for vision problems and that vision problems are an important application domain for machine learning tools.
The question is if there is something you would like in skimage that would require the use of something from sklearn. I.e. what algorithm you want in skimage is only useful together with something from sklearn? I don't think any of your examples are in this category. Which is well enough, since this means it should be easy to keep the two things separate ;) Btw, MRFs in vision are often not learned, so this is no ML, just optimization. And I would rather place that in skimage, as it is quite image specific. When I talked about graph-cuts, that's what I meant. Normalized cuts are of limited use in low level vision since they are very slow for superpixels. I was rather thinking of Boykov-Kolmogorov push-relabel - which is my next project when I get my superpixels done. (Stefan actually mentioned he'd like to have it :) Abour RBMs: I am <http://www.ais.uni-bonn.de/deep_learning/papers/ESANN10_schulz.pdf> no <http://www.ais.uni-bonn.de/deep_learning/papers/IJCNN10_mueller.pdf> stranger <http://www.ais.uni-bonn.de/%7Eschulz/publications/papers/nips10ws_schulz_mueller_behnke.pdf> to them <http://www.ais.uni-bonn.de/%7Eschulz/publications/papers/neurocomp11_schulz_mueller_behnke.pdf> but I would rather not include them in sklearn and definitely not in skimage. Cheers, Andy
Hi Andy: Thanks for your comments. Seems I am not quite clear about the scope of skimage and scikits-learn currently. I will get myself more familiar with both packages to find anything currently in skimage requires the overlap with scikits-learn's methods. For MRF, I think your comments are correct. It is just related to some specific forms(if not one) of MRFs and to solve it is the main problem. Maybe a complete functioning MRF package is not quite necessary and not suitable/efficient enough to do the image tasks. Could I ask why rbm will not be in both packages? I see you release a RBM on your blog, and why not make it into scikits-learn. Seems it is a big-package itself and some python libarary like theano has been relaesed, maybe this is the concern? Best, Wei On Fri, Jul 27, 2012 at 6:02 AM, Andreas Mueller <amueller@ais.uni-bonn.de>wrote:
Hi Wei.
Thanks for your reply! I see your great work in scikits-learn and your comments are quite useful.
Thanks :)
Many will overlap, I think we can maintain a list. As a quick example in the following:
There is no discussion that machine learning methods are helpful for vision problems and that vision problems are an important application domain for machine learning tools.
The question is if there is something you would like in skimage that would require the use of something from sklearn. I.e. what algorithm you want in skimage is only useful together with something from sklearn? I don't think any of your examples are in this category. Which is well enough, since this means it should be easy to keep the two things separate ;)
Btw, MRFs in vision are often not learned, so this is no ML, just optimization. And I would rather place that in skimage, as it is quite image specific. When I talked about graph-cuts, that's what I meant. Normalized cuts are of limited use in low level vision since they are very slow for superpixels. I was rather thinking of Boykov-Kolmogorov push-relabel - which is my next project when I get my superpixels done. (Stefan actually mentioned he'd like to have it :)
Abour RBMs: I am<http://www.ais.uni-bonn.de/deep_learning/papers/ESANN10_schulz.pdf> no <http://www.ais.uni-bonn.de/deep_learning/papers/IJCNN10_mueller.pdf> stranger<http://www.ais.uni-bonn.de/%7Eschulz/publications/papers/nips10ws_schulz_mueller_behnke.pdf>to them<http://www.ais.uni-bonn.de/%7Eschulz/publications/papers/neurocomp11_schulz_mueller_behnke.pdf>but I would rather not include them in sklearn and definitely not in skimage.
Cheers, Andy
participants (2)
-
Andreas Mueller
-
LI Wei