b34b9d8ddd
ENH: Only perform convex hull on select candidate pixels. Gets us a factor 30 speedup.
2011-10-26 23:17:00 -07:00
74f7e01e42
ENH: Improve convex_hull execution speed.
2011-10-26 19:55:46 -07:00
6bbeda04f3
ENH: Pre-allocate coordinate corner array.
2011-10-26 19:37:34 -07:00
b5ad567a68
ENH: Add binary convex hull computation.
2011-10-26 17:43:30 -07:00
573e208361
Merge pull request #67 from NeilYager/neil_yager-alpha
...
BUG: Ignore alpha channel when opening RGBA image as grayscale.
2011-10-25 01:02:53 -07:00
d56ae9631c
Bug fix: Ignore alpha channel when opening RGBA image as grayscale
2011-10-25 08:52:13 +01:00
0b8452fa8e
Merge Emmanuelle's port of CellProfiler's medial axis skeletonization.
2011-10-24 17:24:11 -07:00
f8a51a4e3e
DOC: Minor spacing fixes in docstrings.
2011-10-24 17:23:39 -07:00
39a0e3577c
scikits/image -> skimage in examples
2011-10-24 23:28:21 +02:00
25e986d545
Moved cython file that had not been automaticall moved by rename
...
scikits/image -> skimage
2011-10-24 23:27:29 +02:00
ed1f05427e
ENH address comments on PR #65 made by Stefan and Neil
2011-10-24 23:23:52 +02:00
736802bad4
Medial axis transform (skeletonization) from cell profiler
...
* medial_axis function in morphology.skeletonize
* plot_medial_axis.py example for the gallery
Compared to the skeletonize algorithm in morphology.skeletonize,
medial_axis is faster because it processes all pixel in one pass.
However, the resulting skeleton has more branches, that may be unwanted
depending on the application.
2011-10-24 23:22:58 +02:00
0f31034a46
First integration of cell profiler medial axis transform
...
(skeletonization in cell profiler)
2011-10-24 23:22:58 +02:00
d5120a16bc
DOC: Update gitwash.
2011-10-23 22:56:43 -07:00
5ec4fdec09
PKG: Import skimage instead of scikits.image in changes made since rename.
2011-10-23 22:55:55 -07:00
a76f12f747
DOC: Update repo address.
2011-10-23 22:51:38 -07:00
7542f06e9d
Rename package to `skimage`.
2011-10-23 22:40:11 -07:00
18775195dd
DOC: Add Nelle to contributors.
2011-10-23 22:39:51 -07:00
91d9949282
PKG: Move remaining files that changed since skimage rename.
2011-10-23 22:34:28 -07:00
3ac50d0a43
Merge remote-tracking branch 'nelle/skimage' into nelle_varoquaux_skimage_rename
2011-10-23 22:31:57 -07:00
6c891fd1e1
Merge branch 'neil_yager-skeletonize'
2011-10-22 12:14:27 +02:00
c45db9f4cc
ENH: more efficient binary operation in skeletonize
2011-10-22 12:10:13 +02:00
c48aa63323
Address 2nd round of review comments
2011-10-21 15:27:33 +01:00
1704ab31a9
skimage/version.py should be ignored. Doc has been updated with new module
...
name ``sklearn``
2011-10-18 21:14:21 +02:00
6d67bba941
Renaming: scikits.image -> skimage
2011-10-18 21:12:27 +02:00
114cf44c77
Renamed scikits.image to skimage
2011-10-18 21:12:27 +02:00
b25011514b
Merge pull request #63 from amueller/matplolib_plugin
...
Use matplotlib plugin for I/O if PIL is not available.
2011-10-18 12:05:19 -07:00
03e8a61fb4
COSMIT removed unused import
2011-10-18 16:56:41 +02:00
fb987ac6ed
add qt plugin again.
2011-10-18 16:55:56 +02:00
3ddfdbdb9c
Updates based on review comments
2011-10-17 08:32:19 +01:00
6ac1927518
try to load only plugins that have both imread and imshow
2011-10-16 19:40:28 +02:00
d0041a8c74
[ENH] include matplotlib's imread
2011-10-16 19:40:25 +02:00
9545f85f18
removed debugging code
2011-10-16 19:40:22 +02:00
22a424bc50
this depends on the plugin used
2011-10-16 19:40:19 +02:00
d9a3a46fb2
gracefully ignore absent plugins
2011-10-16 19:40:11 +02:00
0cfc440b5d
Merge pull request #62 from tonysyu/fix_morph_symmetry
...
Fix morph symmetry
2011-10-16 05:46:47 -07:00
1cc0f36d15
Reformat doctests so that tests pass
...
I believe the dtype print out was changed to a string in a recent version of numpy, so this change may actually break the doctest for older versions.
2011-10-15 12:44:26 -04:00
9e81cecbb6
Specify that input arrays to grey morphology filters must be uint8
2011-10-15 12:40:06 -04:00
c7c0798d1c
Add examples to grey morphology docstrings
2011-10-15 12:37:25 -04:00
8735204565
Improve docstrings in grey morphology module
2011-10-15 11:57:01 -04:00
65b6b6bed7
Minor fix to docstring
2011-10-15 04:27:35 -04:00
b0376721eb
Replace flip parameter in erode/dilate with shift_x and shift_y
...
This change allows morphological open/close and white/black tophat to correctly use rectangular structuring element that has one even-numbered dimension. Note: dilate and erode functions flip the definitions of x/y and width/height, but it does so consistently so it doesn't really matter.
2011-10-15 04:23:37 -04:00
6366da861a
Add tests for rectangular structuring elements
...
These tests fail.
2011-10-15 04:19:04 -04:00
9ce5ee7d71
Move test functions to test class
2011-10-15 03:59:41 -04:00
ac2714d3c8
Fix morphological open/close for square structuring elements with even sides.
...
This also fixes rectangular selems with even sides, but not rectangular selems with one odd and one even side.
2011-10-15 03:34:30 -04:00
fb92bfcf71
Add tests for correct behavior of square structuring elements
2011-10-15 03:32:59 -04:00
4f4f219788
Reuse open/close operations for white/black tophat
2011-10-15 03:01:27 -04:00
b6e470dbd1
Remove unnecessary if-block.
...
This looks like it was a copy-paste error.
2011-10-15 02:57:32 -04:00
9adbca13e7
Add parameter to flip structuring element about its origin
...
This parameter only affects structuring elements (selem) with even-numbered sides, since an odd-numbered selem is centered about its origin. Note the selem isn't actually flipped, but shifted (i.e. I assume a selem that is symmetric in *shape*)
2011-10-14 11:00:10 -04:00
fb81057774
Added a new unit test
2011-10-14 12:53:06 +01:00
Stefan van der Walt