mirror/tesseract
1
0
Fork 0
mirror of https://github.com/tesseract-ocr/tesseract.git synced 2024-12-22 06:17:49 +08:00
Code Issues Actions 10 Packages Projects Releases Wiki Activity
4,844 Commits 10 Branches 66 Tags 222 MiB
d1d176a280
Commit Graph

4844 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Shatur95
068b79052f Fix typo 2020-12-11 18:39:13 +02:00
Shatur95
8719498c19 Locate CMake config files without subfolder 2020-12-11 18:24:51 +02:00
Shatur95
30e2516337 Use find_dependency instead of find_package 2020-12-11 18:05:09 +02:00
Stefan Weil
66175f9745
Merge pull request #3172 from MerlijnWajer/master 
tesseract.1.asc: sync with languages available in tessdata-fast
 2020-12-08 06:00:15 +01:00
Stefan Weil
3f2892bc04
Update description for fry language to match Wikipedia 2020-12-08 05:59:17 +01:00
softana
bb832d491e
Update Dockerfile 
Change double hyphen "--" to single hyphen "-" to prevent build errors:
Fix invalid option no-ri-no-rdoc
> ERROR: While executing gem ... (OptionParser::InvalidOption) invalid option: --no-ri
 2020-12-07 11:33:09 -06:00
Stefan Weil
43e13ea6f4
Merge pull request #3171 from stweil/lsan 
Suppress some LeakSanitizer errors in unit tests
 2020-12-05 10:20:54 +01:00
Merlijn Wajer
5ff273675c tesseract.1.asc: sync with languages available in tessdata-fast 
cos, div, fao, fyr, gla, hye are available in Ubuntu's 'tesseract-ocr-*'
packages but not mentioned in the manpage.
 2020-12-04 18:16:45 +01:00
Stefan Weil
b303dd6ac2 Add more patterns to suppress memory leaks from libfontconfig 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-12-04 13:30:58 +01:00
Stefan Weil
490bd3ec8f Fix build with enabled TensorFlow 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-12-04 10:56:23 +01:00
Stefan Weil
5eb5e6ea23 Suppress some LeakSanitizer errors in unit tests 
The fontconfig library has some (intentional) memory leaks which
must be suppressed for unit tests with the LeakSanitizer.

This fixes the issues #3156 and #3157.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-12-04 07:25:49 +01:00
Stefan Weil
ac116d1b28 Fix regression in Network::Serialize (fix issue #3167) 
The regression was caused by a wrong string serialization in
commit 4613738a5e.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-12-03 19:36:58 +01:00
Stefan Weil
69ed480a9a
Merge pull request #3165 from MerlijnWajer/master 
Remove references to "kur" and "tgl", add "fil" to man page
 2020-12-03 13:42:55 +01:00
Merlijn Wajer
58f7a72f00 Remove references to "kur" and "tgl", add "fil" to man page 
"kur" no longer exists, might be named "kur_ara" (the old "kur_ara" is
now "kmr", which is actually Latin) now, but "kur" is not present in
tessdata_fast nor in tessdata_best. [1] [2]

"tgl" (Tagalo) is now named "fil" (Filipino) [3]

[1] https://github.com/tesseract-ocr/langdata/issues/124
[2] https://github.com/tesseract-ocr/tessdata_best/issues/23
[3] https://github.com/tesseract-ocr/langdata/issues/84
 2020-12-01 23:43:50 +01:00
zdenop
a06c61cc90
Merge pull request #3128 from acoder77/patch-1 
Create .gitattributes for cross os contributors
 2020-11-27 18:27:26 +01:00
zdenop
279b0b2e37
Merge pull request #3160 from stweil/string2 
Replace more occurrences of STRING by std::string of char*
 2020-11-27 18:24:17 +01:00
zdenop
6bc42464af
Merge pull request #3159 from stweil/pack 
Pack BlamerBundle, CLASS_STRUCT and SVMenuNode
 2020-11-27 18:23:14 +01:00
Stefan Weil
65b11a1e12 Pack class SVMenuNode 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-11-26 17:17:27 +01:00
Stefan Weil
a1849bc65c Pack struct CLASS_STRUCT 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-11-26 17:17:27 +01:00
Stefan Weil
0bb46ac2e0 Pack struct BlamerBundle 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-11-26 17:17:27 +01:00
Stefan Weil
bf3774cc91 Use more const char* 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-11-26 17:01:17 +01:00
Stefan Weil
4613738a5e Use const char* for filename and network_spec parameters 
This replaces the proprietary STRING data type
(764 instead of 838 lines remaining).

It also removes STRING from osdetect.h and serialis.h.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-11-26 17:01:17 +01:00
Amit D
4c35f51a5c
Merge pull request #3158 from Shreeshrii/master 
fixes issue #3099
 2020-11-25 03:48:42 +02:00
Shree Devi Kumar
31710098e3 fixes issue 3099 2020-11-23 13:30:26 +00:00
Egor Pugin
dea08c34f8
Merge pull request #3155 from Shatur95/fix-cmake-targets-path 
Fix CMake targets path
 2020-11-18 04:26:10 +03:00
Shatur95
80147735db Fix CMake targets path 2020-11-18 02:01:55 +02:00
zdenop
e20ffdd719
Merge pull request #3153 from stweil/scale 
Remove GenericVector::scale() again and replace more STRING by std::string
 2020-11-12 20:01:46 +01:00
Stefan Weil
fbc4c809d9 Replace STRING by std::string 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-31 14:08:39 +01:00
Stefan Weil
92b6c652f3 Use std::vector for scales_ 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-29 08:00:11 +01:00
Stefan Weil
c15dd26b84 Don't pass scales_ to IntSimdMatrix::Init 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-28 20:35:53 +01:00
Stefan Weil
fe76142a3d Remove GenericVector::scale() again 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-28 16:24:59 +01:00
zdenop
5761880676
Merge pull request #3141 from stweil/invert 
Modify OCR for inverted text
 2020-10-27 08:57:21 +01:00
Stefan Weil
eaf72ace31 Prefer result from inverted image if the mean confidence is better 
Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-26 20:37:47 +01:00
Stefan Weil
cfb1fb2540 Try OCR on inverted line only if mean confidence is below 50 % 
The old code looked for the minimum confidence which triggered
very often a 2nd OCR without improving the result.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
 2020-10-26 09:32:09 +01:00
zdenop
11297c983e
Merge pull request #3130 from robinwatts/pushback15 
Tweak SIMDDetect for ANDROID Neon.
 2020-10-19 18:21:00 +02:00
Robin Watts
436008bd37 Tweak SIMDDetect for ANDROID Neon. 
cpufeatures.h should be cpu-features.h, with the latest NDK
at least. The #if 0'd section is not required because armv8
always includes NEON.
 2020-10-19 12:04:29 +01:00
zdenop
514a7893f4
Merge pull request #2994 from robinwatts/pushback11 
Improve speed of tesseract by optimising for intSimdMatrix case
 2020-10-17 17:19:49 +02:00
acoder77
ac661414b5
Create .gitattributes for cross os contributors 
With this set, Windows users will have text files converted from Windows style line endings (\r\n) to Unix style line endings (\n) when they’re added to the repository.
https://www.edwardthomson.com/blog/git_for_windows_line_endings.html
 2020-10-17 11:23:42 +05:30
Robin Watts
db10c7b577 intsimdmatrixneon.cpp: Do biasing in SIMD. 2020-10-12 04:30:46 -07:00
Robin Watts
d1e49d6dd2 intsimdmatrixavx2: Do biasing in SIMD. 
We also move to relying on both scales and output having been
padded to accomodate us writing more results than are actually
needed here. This was allowed for a few commits back.
 2020-10-12 04:30:46 -07:00
Robin Watts
872816897a Rejig intsimdmatrix to reduce FP ops. 
Avoid 1) floating point division by 127, 2) conversion of
bias to double, 3) FP addition, in favour of 1) integer
multiplication by 127, and 2) integer addition.

(Also costs extra work in the serialisation/deserialisation of
the scale values, and conversion of weights to int formats, but
these are all one offs).
 2020-10-12 04:30:46 -07:00
Robin Watts
aba1800f69 Round output buffers for intSimdMatrix. 
In order to allow intSimdMatrix implementations to 'overwrite'
their outputs, ensure that the output buffers are always padded
to the next block size.

This doesn't make any difference yet, but it enables optimisations
further down the line, especially when the biasing is pulled into
the SIMD.
 2020-10-12 11:47:16 +01:00
Robin Watts
9dfdac51c6 Tweak scales array for intSimdMatrix case. 
Currently, the size of the scales array is not rounded up
in the same way as the weights are. This blocks us pushing
the scale calculations into the SIMD, as when we "overread"
the end of the scale array, we potentially get errors.

Here, we adjust the intSimdMatrix stuff to ensure that the
scales array reserves enough entries to allow such overreads
to work.

This doesn't make any difference for now, but opens the way
for future optimisations.
 2020-10-12 11:47:16 +01:00
Shatur95
5a377707e0 Generate imported target automatically 2020-10-12 11:47:16 +01:00
Shatur95
8dad1e24a2 Modernize CMake config files 2020-10-12 11:47:16 +01:00
amitdo
958f23453e Improve disabled legacy engine build 2020-10-12 11:47:16 +01:00
amitdo
06154e028b Improve disabled legacy engine build 2020-10-12 11:47:16 +01:00
amitdo
e81b485066 Improve disabled legacy engine build 2020-10-12 11:47:15 +01:00
amitdo
7df4918644 Improve disabled legacy engine build 2020-10-12 11:47:15 +01:00
Shatur95
ec8766ce74 Use DESTINATION instead of TYPE 
For compatibility with older CMake.
 2020-10-12 11:47:15 +01:00