[SciPy-User] [Numpy-discussion] Simple pattern recognition

[Zachary Pincus]

No worries! I think I've written connected-component finding code  
several times over in different guises. Definitely a good exercise.


On Sep 21, 2009, at 3:14 PM, Gökhan Sever wrote:

> Ahh my blindness and apologies :)
>
> The nice feeling of reinventing the wheel...
>
> Probably I forgot to reshape the image data in the first place  
> before applying into ndimage.label().
>
> However, this was a nice example to understand recursion, and get to  
> know some basics of computer vision and few libraries (OpenCV,  
> pygraph) during my research.
>
> Thanks again for all kind replies.
>
> On Mon, Sep 21, 2009 at 1:36 PM, David Warde-Farley <dwf at cs.toronto.edu 
> > wrote:
> I think Zachary is right, ndimage does what you want:
>
> In [48]: image = array(
> [[0,0,0,1,1,0,0],
> [0,0,0,1,1,1,0],
> [0,0,0,1,0,0,0],
> [0,0,0,0,0,0,0],
> [0,1,0,0,0,0,0],
> [0,1,1,0,0,0,0],
> [0,0,0,0,1,1,0],
> [0,0,0,0,1,1,1]])
>
> In [57]: import scipy.ndimage as ndimage
>
> In [58]: labels, num_found = ndimage.label(image)
>
> In [59]: object_slices = ndimage.find_objects(labels)
>
> In [60]: image[object_slices[0]]
> Out[60]:
> array([[1, 1, 0],
>        [1, 1, 1],
>        [1, 0, 0]])
>
> In [61]: image[object_slices[1]]
> Out[61]:
> array([[1, 0],
>        [1, 1]])
>
> In [62]: image[object_slices[2]]
> Out[62]:
> array([[1, 1, 0],
>        [1, 1, 1]])
>
> David
>
> On 21-Sep-09, at 2:04 PM, Gökhan Sever wrote:
>
> > ndimage.label works differently than what I have done here.
> >
> > Later using find_objects you can get slices for row or column basis.
> > Not
> > possible to construct a dynamical structure to find objects that are
> > in the
> > in both axis.
> >
> > Could you look at the stackoverflow article once again and comment
> > back?
> >
> > Thanks.
> >
> > On Mon, Sep 21, 2009 at 12:57 PM, Zachary Pincus <zachary.pincus at yale.edu
> > >wrote:
> >
> >> I believe that pretty generic connected-component finding is  
> already
> >> available with scipy.ndimage.label, as David suggested at the
> >> beginning of the thread...
> >>
> >> This function takes a binary array (e.g. zeros where the background
> >> is, non-zero where foreground is) and outputs an array where each
> >> connected component of non-background pixels has a unique non-zero
> >> "label" value.
> >>
> >> ndimage.find_objects will then give slices (e.g. bounding boxes)  
> for
> >> each labeled object (or a subset of them as specified). There are
> >> also
> >> a ton of statistics you can calculate based on the labeled  
> objects --
> >> look at the entire ndimage.measurements namespace.
> >>
> >> Zach
> >>
> >> On Sep 21, 2009, at 1:45 PM, Gökhan Sever wrote:
> >>
> >>> I asked this question at
> >> http://stackoverflow.com/questions/1449139/simple-object- 
> recognition
> >>> and get lots of nice feedback, and finally I have managed to
> >>> implement what I wanted.
> >>>
> >>> What I was looking for is named "connected component labelling or
> >>> analysis" for my "connected component extraction"
> >>>
> >>> I have put the code (lab2.py) and the image (particles.png) under:
> >>> http://code.google.com/p/ccnworks/source/browse/#svn/trunk/ 
> AtSc450/
> >>> labs
> >>>
> >>> What do you think of improving that code and adding into scipy's
> >>> ndimage library (like connected_components())  ?
> >>>
> >>> Comments and suggestions are welcome :)
> >>>
> >>>
> >>> On Wed, Sep 16, 2009 at 7:22 PM, Gökhan Sever
> >>> <gokhansever at gmail.com> wrote:
> >>> Hello all,
> >>>
> >>> I want to be able to count predefined simple rectangle shapes on  
> an
> >>> image as shown like in this one:
> >> http://img7.imageshack.us/img7/2327/particles.png
> >>>
> >>> Which is in my case to count all the blue pixels (they are ice- 
> snow
> >>> flake shadows in reality) in one of the column.
> >>>
> >>> What is the way to automate this task, which library or technique
> >>> should I study to tackle it.
> >>>
> >>> Thanks.
> >>>
> >>> --
> >>> Gökhan
> >>>
> >>>
> >>>
> >>> --
> >>> Gökhan
> >>> _______________________________________________
> >>> SciPy-User mailing list
> >>> SciPy-User at scipy.org
> >>> http://mail.scipy.org/mailman/listinfo/scipy-user
> >>
> >> _______________________________________________
> >> SciPy-User mailing list
> >> SciPy-User at scipy.org
> >> http://mail.scipy.org/mailman/listinfo/scipy-user
> >>
> >
> >
> >
> > --
> > Gökhan
> > _______________________________________________
> > SciPy-User mailing list
> > SciPy-User at scipy.org
> > http://mail.scipy.org/mailman/listinfo/scipy-user
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> http://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
>
> -- 
> Gökhan
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> http://mail.scipy.org/mailman/listinfo/numpy-discussion