Coverity report:
CID 1164737 (#1 of 1): Resource leak (RESOURCE_LEAK)
49. leaked_storage: Variable p going out of scope leaks the storage it points to.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
In osdetect.cpp, a local definition of kMinCredibleResolution was
identical to a global one, so the local one could be removed.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Revert fd429c32, 43834da7, 05de195e.
See #49, #59.
The code in this commit solves the issue in a more elegant way, IMHO.
Now you can use:
* `tesseract eurotext.tif eurotext txt pdf`
* `tesseract eurotext.tif eurotext txt hocr`
* `tesseract eurotext.tif eurotext txt hocr pdf`
NOTE:
With `tesseract eurotext.tif eurotext`
or `tesseract eurotext.tif eurotext txt`
the psm will be set to '3', but...
With `tesseract eurotext.tif eurotext txt pdf`
or `tesseract eurotext.tif eurotext txt hocr`
the psm will be set to '1'.
Coverity bug report: CID 1270401 (#1 of 1): Use after free
As the comment (which was also fixed) says, ReadNextBox() already
calls fclose(box_file), so don't call it a 2nd time.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
This fixes compiler warnings like this one:
api/baseapi.h:739:32: warning:
type qualifiers ignored on function return type [-Wignored-qualifiers]
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Font recognition was poor, due to forcing a 1st and 2nd choice at
a character level, when the total score for the correct font is often
correct at the word level, so allowed the propagation of a full set
of fonts and scores to the word recognizer, which can now decide word
level fonts using the scores instead of simple votes.
Change precipitated a cleanup of output data structures for classifier
results, eliminating ScoredClass and INT_RESULT_STRUCT, with a few
extra elements going in UnicharRating, and using that wherever possible.
That added the extra complexity of 1-rating due to a flip between 0 is
good and 0 is bad for the internal classifier scores before they are
converted to rating and certainty.
Tha, Vie, Kan, Tel etc.
There is a new overlap detector that detects when diacritics
cause a big increase in textline overlap. In such cases, diacritics from
overlap regions are kept separate from layout analysis completely, allowing
textline formation to happen without them. The diacritics are then assigned
to 0, 1 or 2 close words at the end of layout analysis, using and modifying
an old noise detection data path.
The stored diacritics are used or not during recognition according to the
character classifier's liking for them.
Eliminated the flexfx scheme for calling global feature extractor functions
through an array of function pointers.
Deleted dead code I found as a by-product.
This CL does not change BlobToTrainingSample or ExtractFeatures to be full
members of Classify (the eventual goal) as that would make it even bigger,
since there are a lot of callers to these functions.
When ExtractFeatures and BlobToTrainingSample are members of Classify they
will be able to access control parameters in Classify, which will greatly
simplify developing variations to the feature extraction process.
a heap checker.
SEAM and SPLIT have been begging for a refactor for a *LONG* time.
This change does most of the work of turning them into proper classes:
Moved relevant code into SEAM/SPLIT/TBLOB/EDGEPT etc from global helper functions.
Made the splits full data members of SEAM in an array instead of 3 separate pointers.
This greatly reduces the amount of new/delete happening in the chopper, which is the main goal.
Deleted redundant files: olutil.*, makechop.*
Brought other code into SEAM in order to keep its data members private with only priority having accessors.
