Commit Graph

69 Commits

Author SHA1 Message Date
Julien Coste 8a337dbe46 Move canny from filter to feature 2014-08-31 10:55:12 +01:00
Johannes Schönberger f9fd683fcb Test missed line for coverage 2014-08-08 19:51:10 -04:00
Johannes Schönberger 23fd7b89af Improve test coverage 2014-08-08 17:01:52 -04:00
Johannes Schönberger ce0e9174c9 Fix test cases which are now wrong due to previous binning error 2014-08-08 17:01:52 -04:00
Juan Nunez-Iglesias 838617cb39 Fix incorrect test function signature 2014-07-28 13:01:41 +10:00
Juan Nunez-Iglesias 278a0d6862 Bug fix: nbins can be >256 in threshold_isodata
The threshold_isodata function created an arange of values up to the
number of bins, but gave it type np.uint8, limiting the number of bins
to 256.

Fixes #1085.
2014-07-28 12:38:34 +10:00
Johannes Schönberger ed558d672e Remove deprecated skimage.filter.median_filter 2014-04-11 10:05:51 -04:00
Johannes Schönberger ef63523ab0 Move denoise functions to restoration submodule 2014-03-13 19:39:54 -04:00
radioxoma cd01bad721 fix tests 2014-01-11 02:00:03 +03:00
radioxoma 4d5889964b Unit-tests and small fixes 2014-01-10 22:32:34 +03:00
radioxoma 2c197846f9 Add ISODATA threshold with tests 2014-01-10 22:32:34 +03:00
radioxoma dfa2ba7bcd fix yen blank image bug 2014-01-10 22:32:34 +03:00
radioxoma 9e74f4eb70 Added Yen tests, typo fixed 2013-09-07 19:10:40 +03:00
Emmanuelle Gouillart 9795d3fad4 Gaussian filter function: changed the default value of multichannel
Now we try to guess automatically whether the image is grayscale or RGB
2013-08-29 23:14:30 +02:00
Emmanuelle Gouillart b4242ca3fb [BUG] iterable sigma and multichannel
+ docstring improvement
2013-08-25 16:41:37 +02:00
Emmanuelle Gouillart 082586c10c Added a wrapper around ndimage's Gaussian filter
This version of the Gaussian filter

 * uses 'nearest' as the default boundary mode. This can be discussed,
   but I had the impression that for images this is the most relevant
   mode ('extending' boundaries)

 * has a `multichannel` keyword, so that each color channel can be
   filtered separately. For now no attempt is made at guessing whether
   the image has color channels or not.
2013-08-25 14:50:55 +02:00
Tony S Yu 5c3c75cd09 Minor cleanups
* PEP8 clean ups (space after commas, docstring spacing, prefer parentheses over line continuation)
* Simplify creation of test images.
* Use `plt.subplots` instead of state machine
* Minor wording modifications
2013-04-25 22:47:22 -05:00
umesh 18839cdf08 Updated Test Case 2013-04-23 08:43:05 -07:00
Umesh 27026cea87 updated tests for Roberts Edge Detector 2013-04-22 07:51:44 -07:00
Umesh e6dedc68e8 Upadated Version with test added(/filter/tests/test_edge.py 2013-04-22 00:21:26 -07:00
Tony S Yu 2838b97261 Merge pull request #371 from ahojnnes/gabor
Gabor filter
2013-04-07 14:12:41 -07:00
Tony S Yu 0f2628c7db Add better test of gabor filter 2013-04-07 10:43:43 -05:00
Tony S Yu 2bf178ceb2 Add bandwidth parameter to gabor_kernel
Note that this changes the API of `gabor_kernel` and `gabor_filter`: The input parameters are rearranged and positional arguments are changed to keyword args.
2013-04-06 21:18:53 -05:00
Tony S Yu 7d407fc778 Fix kernel-size calculation for non-zero theta. 2013-04-06 20:27:05 -05:00
Johannes Schönberger fcc7edd411 Set random seed value for denoising tests 2013-01-17 09:11:21 +01:00
Johannes Schönberger db5460591a Add test case for multi-channel chambolle denoising 2012-12-26 09:30:13 +01:00
Johannes Schönberger 5078da0aed Add test cases for chambolle tv denoising implementation 2012-12-26 09:30:13 +01:00
Johannes Schönberger ae5488a2b1 Rename test functions for bregman algorithm 2012-12-26 09:30:12 +01:00
Johannes Schönberger 780d886db9 Rename denoise_tv to denoise_tv_bregman 2012-12-26 09:28:25 +01:00
Johannes Schönberger cda03cfba4 Fix test cases for TV denoising 2012-12-26 09:28:25 +01:00
Johannes Schönberger aeaf2e1a12 Add test cases for gabor filter 2012-10-28 22:32:28 +01:00
Johannes Schönberger c04f14c129 Combine 2D and 3D bilateral filter in one function and rename sigma parameters 2012-10-17 16:28:44 +02:00
Johannes Schönberger 5b8f554a2e Rename tv_denoise to denoise_tv and deprecate 2012-10-17 12:09:09 +02:00
Johannes Schönberger 7967d5fb49 Refactor denoise tests and add tests for bilateral filter 2012-10-17 09:12:27 +02:00
Johannes Schönberger 3757e5c5eb Use common file for denoising filters 2012-10-17 09:12:27 +02:00
Johannes Schönberger 239026a19d Add scharr filter to mask test cases 2012-10-13 09:21:10 +02:00
Johannes Schönberger 37ccc805e9 Fix test case doc strings 2012-10-13 09:19:53 +02:00
Johannes Schönberger f22a014ffa Add scharr edge filter 2012-10-12 19:10:46 +02:00
Johannes Schönberger 00dbb5e4e3 Improve test coverage of LPI filter 2012-10-06 18:36:15 +02:00
Johannes Schönberger bca962cc8a Improve test coverage for ctmf 2012-10-06 17:59:55 +02:00
Johannes Schönberger 01df23a67d Full test coverage for canny filter 2012-10-06 17:47:54 +02:00
Andreas Mueller 6c59e04714 MISC remove unused imports, some pep8 corrections. 2012-09-27 20:01:45 +01:00
Tony S Yu 852481e055 TST: Add tests for masked region 2012-09-03 22:27:51 -04:00
Tony S Yu 6e3d460b3c STY: Rename tests.
Test classes were unnecessary. Simplify to functions.
2012-09-03 08:08:29 -04:00
Tony S Yu f2246027fd STY: Clean up imports 2012-09-03 07:55:05 -04:00
Andreas Wuerl 12b8d8d051 cleanup of tests hence all results of tv denoise operations are returned as float 2012-09-02 21:17:18 +02:00
Andreas Wuerl 42ae537a69 convert image to float before performing tv_denoise operation now
removed keep_type argument from tv_denoise which becomes obsolete with the previos change
adapted tests
2012-08-27 19:44:30 +02:00
Andreas Wuerl 14d0923959 fixed data of tv_denoise result images with float datatype to be in default range [0.0:1.0] 2012-08-27 12:41:29 +02:00
Matt McCormick 67eb914ef6 TST: Add buffer for threshold tests.
In case conversion from float to integer is imprecise.
2012-07-20 19:09:05 -04:00
Matt McCormick 4b53c92c14 BUG: Fix OTSU thresholding tests with matplotlib IO plugin.
The matplotlib IO returns float arrays from [0, 1], which gives difference
results than a ubyte array.  Explicitly convert to a ubyte array in the tests.
2012-07-20 18:49:44 -04:00