Andrey Kamaev
2a669555de
Merge pull request #414 from asmaloney:remove-unused-vars2
2013-02-02 14:34:47 +04:00
Andy Maloney
3154cdf8ac
Fix subtle bug when src & dst agree on sparsity but have different dimensions
...
Remove unused var "total"
Declare vars as locally as possible
2013-02-01 22:57:22 -05:00
Andy Maloney
bc68dfb4e8
Remove unused vars
2013-02-01 18:09:58 -05:00
Vadim Pisarevsky
efd00238e2
fixed warnings; restored fixed_size parameter in AutoBuffer
2013-01-23 21:47:58 +04:00
Vadim Pisarevsky
dc4d0398f3
converted few more comp. geometry functions to C++
2013-01-22 16:54:31 +04:00
Vadim Pisarevsky
c2241dccc5
converted fitline to C++, fixed Mat::checkVector().
2013-01-21 01:07:47 +04:00
Vadim Pisarevsky
c197a46e7e
removed rarely used fixed_size parameter from AutoBuffer type, added optional AutoBuffer* but to cvarrToMat in order to speedup CvSeq->Mat conversion; finished conversion of convex hull and related functions to C++
2013-01-20 00:58:51 +04:00
Vadim Pisarevsky
457fa52111
convex hull converted to C++; other 2 functions in convhull.cpp are yet to be finished.
2013-01-19 00:26:01 +04:00
Vadim Pisarevsky
e3941d0965
refactored approxpoly
2013-01-17 00:11:08 +04:00
Andrey Kamaev
084b1c7969
Merge branch 2.4
2013-01-11 15:45:23 +04:00
Vadim Pisarevsky
b68df415a9
Merge pull request #135 from nevion:master
2013-01-10 16:43:45 +04:00
Andrey Kamaev
6131a847a2
Fix histogram calculation in equalizeHist
...
issue #2678
2013-01-10 14:48:31 +04:00
Vadim Pisarevsky
e383d39598
fixed bug #2679
2013-01-09 17:53:19 +04:00
Ilya Lavrenov
1bc76813f3
fixed bug connected with SSE2 version of resize with AREA interpolation
2012-12-31 15:35:40 +04:00
Andrey Kamaev
132b5d6435
Merge release 2.4.3.2
2012-12-25 19:07:08 +04:00
Andrey Kamaev
16687a2655
Explicitly use double precision sin & cos in hough implementation
...
On some platforms sin and cos are calculated in single precision resulting in
diversity of results.
2012-12-25 18:05:05 +04:00
Andrey Kamaev
962884cdec
Merge branch 2.4
2012-12-21 17:58:48 +04:00
Andrey Kamaev
b35fa6c4ff
Merge pull request #202 from Daniil-Osokin:calcHist
2012-12-19 15:07:40 +04:00
Vadim Pisarevsky
31857082f4
probably fixed build problems on Android
2012-12-18 14:12:07 -08:00
Daniil Osokin
98d7d99244
Add threaded version of equalizeHist
2012-12-18 23:59:50 +04:00
Ilya Lavrenov
a319bae6e2
added #undef _mm_packus_epi32
2012-12-17 16:56:37 +04:00
Ilya Lavrenov
d246b41573
changed loop condition
2012-12-17 15:50:01 +04:00
Ilya Lavrenov
c2f2e33a5e
removed debug messages
2012-12-17 15:45:20 +04:00
Ilya Lavrenov
743dce6a4b
Merge remote-tracking branch 'upstream/master' into SIMDFastAreaResize-2
2012-12-17 15:27:32 +04:00
Ilya Lavrenov
aa0dafcc1f
replaced _mm_packs_epi32 to _mm_packus_epi32
2012-12-17 14:09:25 +04:00
Daniil-Osokin
7d94236c14
TBB version of calcHist
2012-12-17 11:01:31 +04:00
Jason Newton
e1b60aa4e3
drop usage of macros... the type is already there!
2012-12-16 22:44:36 -08:00
Jason Newton
5f86eb8d63
proper dimensions from cv::Size in output stats/centroid
2012-12-16 18:33:50 -08:00
Vadim Pisarevsky
1eae455acb
fixed build warnings from VS; hopefully, fixes Android build too
2012-12-15 23:14:50 +04:00
Vadim Pisarevsky
2a42960ff2
Merge branch 'master' of https://github.com/nevion/opencv into cc
2012-12-15 21:45:55 +04:00
Jason Newton
ad0bfdfb25
disable windows build warning for connectedcomponents template argument comparisons
2012-12-15 05:56:18 -08:00
Jason Newton
68e77ac051
use opencv's integer type convension
2012-12-15 04:25:55 -08:00
Andrey Kamaev
0bbba847a4
Fix equalization formula in equalizeHist function & rewrite in C++
...
Old implementation did
lut[i] = 255 * (count(Y <= i)) / (width * height)
which actually shifts uniform histograms.
From now histogram is equalized as
C = count(Y == min(Y))
lut[i] = 255 * (count(Y <= i) - C) / (width * height - C)
2012-12-15 15:29:15 +04:00
Ilya Lavrenov
6059a6875a
fixed bug with s1 calculating
2012-12-14 17:54:07 +04:00
Ilya Lavrenov
4ccb5a30d9
fixed bug with s1 calculating
2012-12-14 17:41:42 +04:00
Ilya Lavrenov
16f9b6f5e4
reproducing C++ version of resize area fast
2012-12-14 14:32:00 +04:00
Ilya Lavrenov
d1ca934115
sse2 version of resize area fast for types CV_(8, 16)UC(1, 3, 4)
2012-12-11 15:14:50 +04:00
Ilya Lavrenov
615e7b2747
added SIMD optimization of Edge-Aware Demosaicing in case of CV_8U
2012-12-10 13:29:08 +04:00
Jason Newton
d094e4bdbe
drop support for 8bit output for size cost relative to utility
2012-12-08 23:09:17 -08:00
Jason Newton
e70b3ef598
use a ltype parameter to determine result Label image type; export stats with differening types over different outputarrays
2012-12-08 23:09:14 -08:00
Ilya Lavrenov
c09a325d3e
fixed error connected with incorrect range
2012-11-30 13:26:40 +04:00
Ilya Lavrenov
98031a4147
removed const modificators
2012-11-30 12:31:33 +04:00
Ilya Lavrenov
b8b13ccd5c
parallel version of Bayer2Gray
2012-11-27 15:27:58 +04:00
Jason Newton
00bdca7684
A few changes to comply with upstream requirements for merge.
...
-Change input/output order from (out Labeled, in Image) -> (in Image, Out Labeled) and convert
to Input/OutputArrays in the process.
-Adopt OutputArray for statistics export so that the algorithm is "wrapper friendly" and not requiring a new struct in
language bindings at the expense of using doubles for everything and slowing statistics computation down..
2012-11-27 02:31:38 -08:00
Ilya Lavrenov
ce5e9a71b5
fixed some warnings on Windows and added debug messages
2012-11-27 13:54:11 +04:00
Ilya Lavrenov
3fa630639f
fixed bayer2bgra in case when size.width <= 2
2012-11-26 19:45:56 +04:00
Ilya Lavrenov
5225672dc0
added edge-aware demosaicing and bayer to bgra conversion
2012-11-26 16:17:43 +04:00
Jason Newton
d5aa679d3f
adjust output type to return int32... it should at least be unsigned but this breaks python bindings;
...
remove non-8bit input type support, not worth the binary size
2012-11-22 21:26:52 -08:00
Jason Newton
85880397c4
connectedcomponents: use opencv integral types, add to docs, fix up things for a python export
2012-11-05 12:02:53 -08:00
Jason Newton
4c0cb2576d
connectedComponents: peep-hole optimizations, mostly surrouding the fact that cv::Mat::at is expensive in a tight-loop -also added a "blobstats" version
2012-11-05 08:10:35 -08:00
