6c67738808
Miscellaneous fixes and improvements
2013-10-12 14:13:10 +02:00
e68ba0db01
Use libc.math.atan2
2013-10-12 14:09:52 +02:00
4f58c076d4
FIX: import M_PI
2013-10-12 14:09:52 +02:00
003dbe419d
MINOR: add missing type
2013-10-12 14:09:52 +02:00
1236e97ff3
MAINT: np.pi -> M_PI
2013-10-12 14:09:52 +02:00
df2ee4d636
remove heapq
2013-10-12 14:09:52 +02:00
e27b798ffa
FIX: handle correctly main axis def
2013-10-12 14:09:51 +02:00
1a9d5bb4ca
TEST: revert precision
2013-10-12 14:09:51 +02:00
594a228a06
TEST: fix bad subs
2013-10-12 14:09:51 +02:00
742699c570
TEST: fix precision
2013-10-12 14:09:51 +02:00
debd4d54d6
ENH: use heapq to select the best match
2013-10-12 14:09:51 +02:00
7e970b18cd
MAINT change HT return API
2013-10-12 14:09:51 +02:00
7a1b1c28de
FIX: fix angle convention
2013-10-12 14:09:51 +02:00
e9f3bd66ac
TEST: fix mistake
2013-10-12 14:09:51 +02:00
5c73432101
Merge pull request #770 from cgohlke/patch-3
...
BUG: Fix ValueError: Buffer dtype mismatch, expected 'long' but got...
2013-10-11 00:13:53 -07:00
4d46bc0912
TST: Fix ValueError: Buffer dtype mismatch, expected 'intp_t' but got 'long'
2013-10-10 21:51:56 -07:00
458556723d
BUG: Fix ValueError: Buffer dtype mismatch, expected 'long' but got 'long long'
...
on win-amd64
2013-10-10 21:48:29 -07:00
ee24f1c2c5
Improve example of warp function to show all available options
2013-10-02 17:45:05 +02:00
40f5e78353
Add support for matrix as inverse_map
2013-10-02 17:33:17 +02:00
9741bfda15
Add note about internal usage of transformation matrix
2013-10-02 17:17:03 +02:00
beeb597ddf
Only clip for ndimage.map_coordinates case
2013-10-02 17:10:46 +02:00
c01e7c9b10
Fix various doc strings using the warp function
2013-09-28 14:39:40 +02:00
a17a1395c3
Fix doc string of warp functions
2013-09-28 14:25:45 +02:00
0ed1e22034
Correcting bug in SimialarityTransform
2013-09-19 10:23:03 +05:30
5f82c206b4
Fix bugs in usage of typed memoryviews
2013-08-19 19:31:40 +02:00
c8f619e384
Use typed memoryviews in transform package
2013-08-19 19:22:28 +02:00
be14f0b28c
Merge pull request #654 from josteinbf/radon-square-center
...
Define rotation axis center for Radon transforms with inverses.
2013-08-06 02:04:32 -07:00
6dd35c8166
Merge pull request #641 from sciunto/example_ht
...
Example Hough transform (ellipse)
2013-08-05 12:22:57 -07:00
a3bb1098d1
Merge pull request #632 from ahojnnes/local-blocks
...
Refactor N-dimensial array resampling and add additional functionality
2013-07-31 14:10:18 -07:00
109f500d4b
Replace deprecated import of _local_func with block_reduce
2013-07-25 22:15:23 +02:00
d5b72f91ab
iradon: Do not suppress 0 frequency for filter=None.
2013-07-14 17:40:17 +02:00
4549204507
iradon: Clean up filter code.
2013-07-14 17:39:48 +02:00
60444ee4d3
Radon transform: PEP8 fixes.
2013-07-13 20:32:15 +02:00
462173a53a
Radon transform: Include boundary in reconstruction circle.
...
A test criterion needed to be relaxed slightly to have tests still
passing. This is ok, as the reconstruction circle is now larger, meaning
larger errors should be expected. Moreover, the test in question uses
random data, and changing the seed causes greater changes in accuracy
than the amount the test criterion was relaxed by.
2013-07-13 20:32:15 +02:00
288ee69483
radon: Correct docstring of order_angles_golden_ratio.
2013-07-13 20:32:15 +02:00
41c6f6d740
radon: Reduce duplication; simplifications.
2013-07-13 20:32:13 +02:00
47b6d0c5a6
Radon transform: Document rotation axis location.
2013-07-13 20:31:28 +02:00
2bfaafd9e2
Radon transform: Redefine projection center for sinograms.
...
This definition is chosen because it is simple to express in the
documentation. No changes in accuracy are to be expected, but comparing
sinograms and reconstructions before and after this commit will give
different results in the cases where ``circle=False`` for ``radon`` or
``iradon``.
2013-07-13 19:31:43 +02:00
4d307f5315
iradon_sart: Fix docstring typo.
2013-07-09 11:44:28 +02:00
1cc42bd943
iradon_sart: Reformat references.
2013-07-09 11:44:03 +02:00
18a06a0add
iradon_sart: Format docs correctly.
2013-07-08 14:50:06 +02:00
c2193cbc34
fix doctest
2013-07-07 10:42:32 +02:00
bea50aa608
iradon: use util.pad for sinogram padding.
2013-07-06 20:55:54 +02:00
b3746b9090
iradon: Cleanup by locating related code in one place.
2013-07-06 20:55:54 +02:00
643a1bc1f5
iradon: Add test for cubic interpolation.
2013-07-06 20:55:53 +02:00
b90ba783ef
iradon: Only allow interpolation methods working well.
...
Of the interpolation methods provided by scipy.interpolate.interp1d,
only cubic has been found to work well with the tests in skimage (other
methods are either identical to linear or nearest, or they produce bad
reconstructions).
2013-07-06 20:55:53 +02:00
2be327815e
radon: Use numpy.interp/scipy.interpolate.
...
Scipy supports all interpolation kinds (nearest, linear) we need, while
numpy supports only linear interpolation. The numpy interpolation is
substantially faster, so this is used even though it complicates the
code slightly.
2013-07-06 20:55:53 +02:00
5955e4e612
iradon: Reduce code duplication.
2013-07-06 20:55:53 +02:00
ac460a9777
iradon: Use util.pad for array padding.
...
This depends on util.pad accepting 0s in pad_width, which has a proposed
solution in gh-634.
2013-07-06 20:55:53 +02:00
237fb989b0
radon: Use util.pad for array padding.
2013-07-06 20:55:53 +02:00
Johannes Schönberger