Mailman 3 November 2014 - scikit-image

Re: 3d blob detection
by Juan Nunez-Iglesias Nov. 26, 2014

Nov. 26, 2014

Hi Joel, I'm the self-appointed Guardian of the of the Dimensions here at scikit-image. =) This is a good idea and I will pop it near the top of the list for conversion to 3D. Of course, pull requests are very welcome! If you would like to try your hand at it, I'm happy to help you through any hairy bits. =) Juan. On Thu, Nov 27, 2014 at 5:42 AM, Thomas Caswell <tcaswell(a)gmail.com> wrote: > It might be worth having a look at > https://github.com/soft-matter/trackpy/pull/162 > > > On Wed Nov 26 2014 at 1:34:54 PM Joel Bernier <joel.bernier(a)gmail.com> > wrote: > >> Any hope of modifying some of the tools in skimage.feature (say >> blob_log() >> <http://scikit-image.org/docs/dev/api/skimage.feature.html#skimage.feature.b…>) >> to handle 3d images? I have a good application in the really of >> processing X-rya diffraction data... >> >> -J >> >> -- >> 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

3d blob detection
by Joel Bernier Nov. 26, 2014

Nov. 26, 2014

Any hope of modifying some of the tools in skimage.feature (say blob_log() <http://scikit-image.org/docs/dev/api/skimage.feature.html#skimage.feature.b…>) to handle 3d images? I have a good application in the really of processing X-rya diffraction data… -J

2 1

regionprops Very slow on centroid identification
by jeff witz Nov. 26, 2014

Nov. 26, 2014

Hello. I'm developing a video extensometer based on the identification of center of mass of circular white mark on a black rubber specimen. In order to deal with data in real time I have to be fast (over 100 fps). So I first identify the Zones Of Interests using this example : http://scikit-image.org/docs/dev/auto_examples/plot_label.html Then I compute the center of mass on each ZOI. As I have a fast camera the ZOI between two successive images doesn't change much. So if I extend the bounding box of my current ZOI I could be pretty confident in the fact that given circular mark in the next picture will be in the extended ZOI and the recompute an updated extened ZOI for the next image. So this is the big picture. You will find bellow the function I use in order to get it : def barycenter(image_,minx_,miny_,maxx_,maxy_,thresh_,border_): bw_=image_[minx_:maxx_+1,miny_:maxy_+1]>thresh_ [Y,X]=np.meshgrid(range(miny_,maxy_+1),range(minx_,maxx_+1)) region=regionprops(bw_) minx,miny,maxx,maxy=region[0].bbox Px_=(X*bw_).sum().astype(float)/bw_.sum() Py_=(Y*bw_).sum().astype(float)/bw_.sum() minx_=X[minx,miny]-border_ miny_=Y[minx,miny]-border_ maxx_=X[maxx,maxy]+border_ maxy_=Y[maxx,maxy]+border_ return Px_,Py_,minx_,miny_,maxx_,maxy_ As you can see I don't use region[0].centroid. I compute the moment myself if I time my function on a 141x108 ZOI I get 504 µs If I time this function : def barycenter2(image_,minx_,miny_,maxx_,maxy_,thresh_,border_): bw_=image_[minx_:maxx_+1,miny_:maxy_+1]>thresh_ [Y,X]=np.meshgrid(range(miny_,maxy_+1),range(minx_,maxx_+1)) region=regionprops(bw_) Px_,Py_=region[0].centroid Px_+=minx_ Py_+=miny_ minx,miny,maxx,maxy=region[0].bbox minx_=X[minx,miny]-border_ miny_=Y[minx,miny]-border_ maxx_=X[maxx,maxy]+border_ maxy_=Y[maxx,maxy]+border_ return Px_,Py_,minx_,miny_,maxx_,maxy_ I get 10ms per loop ! What is really strange is if I time : %timeit region[0].centroid I get 58.6 ns per loop ! So I don't really understand why this time explose when I use it in a function ? If someone have some insight it will be very helpfull. Even If I can use my first function, it's a pity to have to use less optimized functions. Best regards.

4 8

Re: Sobel
by Stéfan van der Walt Nov. 20, 2014

Nov. 20, 2014

Nice suggestion, Tony.

1 0

Re: Sobel
by Johannes Schoenberger Nov. 20, 2014

Nov. 20, 2014

What is the reference implementation in this case? Scipy? > On Nov 19, 2014, at 10:43 PM, Juan Nunez-Iglesias <jni.soma(a)gmail.com> wrote: > > Is it what the "reference" implementation returns? Either way the sign is valuable info. +1 for returning the actual filter output. > > On Thu, Nov 20, 2014 at 11:25 AM, Johannes Schoenberger <jsch(a)demuc.de> wrote: > Hi, > > Why are we returning the absolute values in the hsobel and vsobel functions? > > I do not see the benefit. > > Johannes > > -- > 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.

2 2

Re: Easy masking of RGB images
by Stefan van der Walt Nov. 20, 2014

Nov. 20, 2014

Hi Adam On 2014-11-19 00:01:06, Adam Hughes <hughesadam87(a)gmail.com> wrote: > img1 + img 2 --> adds values > img1 U img2 --> replaces values > > Essentially, overlay all the pixels that are not black. > > Is there a library or set of functions somewhere that makes this easy? I > remember a while back thinking I would have to write my own overlay/masking > utilities to get this behavior, but I hope that's not true. It should be fairly straightforward, if I understand correctly: mask = (img2 != 0) img1[mask] = img2[mask] Regards Stéfan

2 1

Error with importing canny (filter/feature)
by Georges H Nov. 19, 2014

Nov. 19, 2014

Hi everyone, So the problem is that canny is not working under feature while i see all the tutorials online using the feature module if i look at the directory, it s just not there >>> print dir(skimage.feature) ['BRIEF', 'CENSURE', 'ORB', '__all__', '__builtins__', '__doc__', '__file__' , '__name__', '__package__', '__path__', '_daisy', '_hessian_det_appx', '_hog', '_texture', 'blob', 'blob_dog', 'blob_doh', 'blob_log', 'brief', 'brief_cy', 'censure', 'censure_cy', 'corner', 'corner_cy', 'corner_fast', 'corner_foerstner', 'corner_harris', 'corner_kitchen_rosenfeld', 'corner_moravec', 'corner_orientations', 'corner_peaks', 'corner_shi_tomasi' , 'corner_subpix', 'daisy', 'greycomatrix', 'greycoprops', 'hessian_matrix', 'hessian_matrix_det', 'hessian_matrix_eigvals', 'hog', 'local_binary_pattern', 'match', 'match_descriptors', 'match_template', 'orb', 'orb_cy', 'peak', 'peak_local_max', 'plot_matches', 'structure_tensor', 'structure_tensor_eigvals', 'template', 'texture', 'util'] looked up a bit on the web this issue, found that it s in feature, proven by : >>> print dir(skimage.filter) ['LPIFilter2D', '__all__', '__builtins__', '__doc__', '__file__', '__name__' , '__package__', '__path__', '_canny', '_gabor', '_gaussian', '_rank_order', 'canny', 'denoise_bilateral', 'denoise_tv_bregman', 'denoise_tv_chambolle', 'deprecated', 'edges', 'gabor_filter', 'gabor_kernel', 'gaussian_filter', 'hprewitt', 'hscharr', 'hsobel', 'inverse', 'lpi_filter', 'prewitt', 'rank', 'rank_order', 'restoration', 'roberts', 'roberts_negative_diagonal', 'roberts_positive_diagonal', 'scharr', 'sobel', 'threshold_adaptive', 'threshold_isodata', 'threshold_otsu', 'threshold_yen', 'thresholding', 'vprewitt', 'vscharr', 'vsobel', 'wiener'] now the thing is can i fix this and why is it there ? Appreciate any help on the matter. Note : "from skimage import filter" works perfectly

2 4

Easy masking of RGB images
by Adam Hughes Nov. 18, 2014

Nov. 18, 2014

Hi, I often find myself wanting to mask two overlay two color images, where the second image will replace values on the first image rather than adding. IE: img1 + img 2 --> adds values img1 U img2 --> replaces values Essentially, overlay all the pixels that are not black. Is there a library or set of functions somewhere that makes this easy? I remember a while back thinking I would have to write my own overlay/masking utilities to get this behavior, but I hope that's not true. Thanks

1 0

Re: ANN: imageio v1.0
by Almar Klein Nov. 15, 2014

Nov. 15, 2014

> What I'm looking for in an IO package > is proper treatment of less-common data types that are nonetheless > common in scientific imaging, such as uint16 and 3D tiffs. (Prior to > Steven's PRs, skimage did a poor job with those.) Can you comment on that? Imageio supports TIFF through the freeimage plugin, but this is more the common kind of TIFF, and probably not sufficient for scientific TIFF formats. It will need some testing to see what is supported and what not. The best approach would probably be to refactor existing code that can deal with such Tiffs into an imageio plugin. There's C. Gohlke's tifffile, tiffany, pylibtif. I am not familiar enough with TIFF to say which of these sufficiently covers scientists needs (if any). But if either of them is good enough, it should be relatively easy to include it in imageio. - Almar > > Thanks for the post, and for the library! — I think it will lead to good > places. =) > > Juan. > > On Fri, Nov 14, 2014 at 9:40 AM, Almar Klein <almar.klein(a)gmail.com > <mailto:almar.klein@gmail.com>> wrote: > > > Thanks for the announcement! Can you expand on how you see the > eco-system of scikit-image, imread and imageio interacting? > > > Since imageio has a pretty good plugin system, it could be a good > place to implement io for all the weird scientific file formats that > exist. Next to imread(), volread(), etc. imageio also provides the > read() function, which returns a reader object specific to the > plugin. This reader object can have additional methods and may yield > other data then just 2D/3D numpy arrays if this makes sense for the > format. > > So I suppose I am hoping for skimage devs to embrace imageio and > help move it further. > > With that, I suppose skimage could focus on the processing rather > than the io, since ideally, we'd have one library for processing > images, and one for reading/saving images. > > <rant> > Developers tend to have a fear for dependencies. I understand this, > and I try to minimize dependencies myself, but reading/writing > images is a task that warrants a standalone package IMO. It troubles > me a bit that different libraries are each rolling their own image > io functionalities. > </rant> > > As for the imread package: Zach and I considered it when discussing > plans for imageio, but imread is based on C++, which is harder to > maintain and makes it a harder dependency. imageio is pure Python > with ctypes bindings to some libs that are auto-downloaded. > > - Almar > > PS: some fancy stuff that imageio can do which might be interesting > for skimage users: stream images from webcam, export visualizations > to animated gif or mp4. > > > -- > 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 > <mailto:scikit-image%2Bunsubscribe@googlegroups.com>. > For more options, visit https://groups.google.com/d/__optout > <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 > <mailto:scikit-image+unsubscribe@googlegroups.com>. > For more options, visit https://groups.google.com/d/optout.

1 0

some broken links in example notebooks
by Michael Aye Nov. 12, 2014

Nov. 12, 2014

I guess you create the example notebooks somehow automatically with sphinx? Because the descriptive text is written in html which mostly shows up well in the notebook, but the links to other skimage functions do show up weird/broken. I even don't know if it's possible or reasonable to fix that somehow for a locally downloaded notebook, but I just wanted to inform you in case you were not aware. My example is for http://scikit-image.org/docs/dev/auto_examples/plot_blob.html The way it's shown in my notebook (ipython 2.3) can be seen here: https://www.dropbox.com/s/ptzoqfx22yhn5vu/Screenshot%202014-10-08%2017.54.4… If you think it's worth a GH issue, just say the word and I create one. Regards, Michael

2 1