Mailman 3 October 2018 - scikit-image

回复： Comparison of watershed implementations
by imagepy＠sina.com 29 Oct '18

29 Oct '18

I found skimage's watershed is very slow, So I wrote one here:https://github.com/Image-Py/imagepy/blob/master/imagepy/ipyalg/hydrolo… it is only for 8-bit image(or int image), but I think we can do a histogram equalization, than convert to int image, do watershed. the performance is OK. a 3d image, which skimage run more than 20 minutes and crash down, my algrthm cost 30s. ----- 原始邮件 ----- 发件人：Stefan van der Walt <stefanv(a)berkeley.edu> 收件人：scikit-image <scikit-image(a)python.org> 主题：[scikit-image] Comparison of watershed implementations 日期：2018年10月27日 05点48分 I wish the authors also compared accuracy, but here are their performance benchmarks. No surprises :) https://www.mdpi.com/2313-433X/4/10/123/pdf "An Overview of Watershed Algorithm Implementations in Open Source Libraries", J. Imaging 2018, 4(10), 123; https://doi.org/10.3390/jimaging4100123, Anton S. Kornilov and Ilia V. Safonov. Best regards, Stéfan _______________________________________________ scikit-image mailing list scikit-image(a)python.org https://mail.python.org/mailman/listinfo/scikit-image

1 0

Comparison of watershed implementations
by Stefan van der Walt 26 Oct '18

26 Oct '18

I wish the authors also compared accuracy, but here are their performance benchmarks. No surprises :) https://www.mdpi.com/2313-433X/4/10/123/pdf "An Overview of Watershed Algorithm Implementations in Open Source Libraries", J. Imaging 2018, 4(10), 123; https://doi.org/10.3390/jimaging4100123, Anton S. Kornilov and Ilia V. Safonov. Best regards, Stéfan

1 0

回复：Re: Integrate snake method in ImagePy
by imagepy＠sina.com 13 Oct '18

13 Oct '18

Hi Stefan: Thanks for indicate the spell mistake. I had modified it. You mean the document url which I ask for help to translate? I need help indeed, but ImagePy has upgrade a lot, to make it more flexable and powerful. I am sorry, I have no energy to synchronize the document... Now ImagePy can operate both image and table data, (based on np.ndarray or pd.dataframe), It looks like a mixed blood of ImageJ and SPSS. Now ImagePy can be extended in many ways: filter: filter a image slice by slice. simple: operate the 3d image stack or element of the imageplus, such as the lookuptable, roi, free: can run depandding on nothing, such as open a image, close the software. macros: record and run commands again table: integrate method based on pandas.DataFrame widget: some ui element, such as the histogram and nevigater on the right. markdown: show some markdown document tool: a mouse tool, such as a pen, roi...And It can show data as image, table, markdown, chart, or 3dview.Writing a plugin is no more hard than writing a notebook demo, (less than 10 lines mostly, except tool and widget, because we must handle the mouse event or write ui code) And I had wrote some plugins set to solve some specific problem, and earn some money from these projects to support ImagePy's Development, Such as this project to treat the SeaIce image. About the document, It is very important, It is for ui-user, may be a no-programmer, So It meaningless to write a simple api document, It should be a manual tutorial with many picture. So I am writting a demo plugins set called IBook, It is a macros set with many demo image to show some processing skill, It is a tutorial to fresh man. you can use ImagePy's menu Plugins > Install > Install Plugins, then paste "https://github.com/Image-Py/ibook.git" to install it. (Later I want to write a book for it, then I need help to translate) Best RegardYXDragon ----- 原始邮件 ----- 发件人：Stefan van der Walt <stefanv(a)berkeley.edu> 收件人：imagepy(a)sina.com, "Mailing list for scikit-image (http://scikit-image.org)" <scikit-image(a)python.org> 主题：Re: [scikit-image] Integrate snake method in ImagePy 日期：2018年10月13日 03点26分 On Sat, 13 Oct 2018 02:48:10 +0800, imagepy(a)sina.com wrote: > I saw in scikit-image 14, add some snake method (chan_vese, morphological_chan_vese ...), It is very useful, but we need a good interactive enviroment to maxmize theirs advantage. So I integrate them into ImagePy, a plugin framework. Now It supports chan_vese, snake from bound, or snake from a roi. (If you like, it can save a gif animate). > Now ImagePy has integrated most function in scipy.ndimage and scikit-image. And It is esay to integrate any function based on Numpy ndarray. > Some snapshot in attachment. Very cool! Can you remind us of the project URL and documentation links again? Tiny spelling mistake I noticed: the method is called Otsu. Stéfan

1 0

Integrate snake method in ImagePy
by imagepy＠sina.com 12 Oct '18

12 Oct '18

Hi Everyone: I saw in scikit-image 14, add some snake method (chan_vese, morphological_chan_vese ...), It is very useful, but we need a good interactive enviroment to maxmize theirs advantage. So I integrate them into ImagePy, a plugin framework. Now It supports chan_vese, snake from bound, or snake from a roi. (If you like, it can save a gif animate). Now ImagePy has integrated most function in scipy.ndimage and scikit-image. And It is esay to integrate any function based on Numpy ndarray. Some snapshot in attachment. Best RegardYXDragon

2 1

Announcement: scikit-image 0.14.1!
by Juan Nunez-Iglesias 02 Oct '18

02 Oct '18

Announcement: scikit-image 0.14.1 ================================= Hi all, We're happy to announce the release of scikit-image v0.14.1! scikit-image is an image processing toolbox for SciPy that includes algorithmsfor segmentation, geometric transformations, color space manipulation, analysis, filtering, morphology, feature detection, and more. This is our first release under our Long Term Support for 0.14 policy. As areminder, 0.14 is the last release to support Python 2.7, but it will beupdated with bug fixes and popular features until January 1st, 2020. Wheels are available right now on PyPI for Windows, Mac, and Linux. Conda packages will be available soon on conda-forge. This release contains the following changes from 0.14.0: Bug fixes --------- - ``skimage.color.adapt_rgb`` was applying input functions to the wrong axis (#3097) - ``CollectionViewer`` now indexes correctly (it had been broken by an update to NumPy indexing) (#3288) - Handle deprecated indexing-by-list and NumPy ``matrix`` from NumPy 1.15 (#3238, #3242, #3292) - Fix incorrect inertia tensor calculation (#3303) (Special thanks to JP Cornil for reporting this bug and for their patient help with this fix) - Fix missing comma in ``__all__`` listing of ``moments_coord_central``, so it and ``moments_normalized`` can now be correctly imported from the ``measure`` namespace (#3374) - Fix background color in ``label2rgb(..., kind='avg')`` (#3280) - Fix an UnboundLocalVariable error when an image consisting of all NaNs was passed to ``filters.threshold_li`` (#3402) Enhancements ------------ - "Reflect" mode in transforms now works fine when an image dimension has size 1 (#3174) - ``img_as_float`` now allows single-precision (32-bit) float arrays to pass through unmodified, rather than being up-converted to 64-bit (#3110, #3052, #3391) - Speed up rgb2gray computation (#3187) - The scikit-image viewer now works with different PyQt versions (#3157)- The ``cycle_spin`` function for enhanced denoising works single- threaded when dask is not installed now (#3218) - scikit-image's ``io`` module will no longer inadvertently set the matplotlib backend when imported (#3243) - Fix deprecated ``get`` keyword from dask in favor of ``scheduler`` (#3366)- Add missing ``cval`` parameter to threshold_local (#3370) API changes ----------- - Remove deprecated ``dynamic_range`` in ``measure.compare_psnr`` (#3313) Documentation ------------- - Improve the documentation on data locality (#3127) - Improve the documentation on dealing with video (#3176) - Update broken link for Canny filter documentation (#3276) - Fix incorrect documentation for the ``center`` parameter of ``skimage.transform.rotate`` (#3341) - Fix incorrect formatting of docstring in ``measure.profile_line`` (#3236) Build process / development --------------------------- - Ensure Cython is 0.23.4 or newer (#3171) - Suppress warnings during testing (#3143) - Fix skimage.test (#3152) - Don't upload artifacts to AppVeyor (there is no way to delete them) (#3315)- Remove ``import *`` from the scikit-image package root (#3265) - Allow named non-core contributors to issue MeeseeksDev commands (#3357, #3358) - Add testing in Python 3.7 (#3359) - Add license file to the binary distribution (#3322) - ``lookfor`` is no longer defined in ``__init__.py`` but rather imported to it (#3162) - Add ``pyproject.toml`` to ensure Cython is present before building (#3295)- Add explicit Python version Trove classifiers for PyPI (#3417) - Ignore known test failures in 32-bit releases, allowing 32-bit wheel builds (#3434) - Ignore failure to raise floating point warnings on certain ARM platforms (#3337) - Fix tests to be compatible with PyWavelets 1.0 (#3406) Credits ------- Made with commits from (alphabetical by last name): - François Boulogne - Genevieve Buckley - Sean Budd - Matthias Bussonnier - Sarkis Dallakian - Christoph Deil - François-Michel De Rainville - Emmanuelle Gouillart - Yaroslav Halchenko - Mark Harfouche - Jonathan Helmus - Gregory Lee - @Legodev - Matt McCormick - Juan Nunez-Iglesias - Egor Panfilov - Jesse Pangburn - Johannes Schönberger - Stefan van der Walt - Hugo VK Reviewed by (alphabetical by last name): - François Boulogne - Emmanuelle Gouillart - Mark Harfouche - Juan Nunez-Iglesias - Egor Panfilov - Stéfan van der Walt - Josh Warner And with the special support of [MeeseeksDev] (https://github.com/MeeseeksBox),created by Matthias Bussonnier

1 0