On Fedora first make sure you have Tesseract OCR installed (and romanian langpack):
sudo yum install tesseract tesseract-langpack-ron
Download and extract tesseract-gui from Sourceforge (version 2.8 was latest at the time of this writting and the patch below applies to this version):
http://sourceforge.net/projects/tesseract-gui/files/tesseract-gui/tesseract-gui-2/
Here is the romanian.patch:
--- tesseract-gui.py 2012-03-20 14:20:37.000000000 +0200
+++ tesseract-gui.py.ro 2012-11-29 19:55:19.721193111 +0200
@@ -1321,6 +1321,7 @@
self.LCdutch:"nld",\
self.LCgerman:"deu",\
self.LCfrench:"fra",\
+ self.LCromanian:"ron",\
self.LCenglish:"eng"}
self.ListLanguages={}
nnfin=len(Languages)
@@ -1329,7 +1330,7 @@
for lang in Languages.keys():
#print Languages[lang]
#print os.path.join(TESSDATA_PATH, Languages[lang] + ".lm")
- if os.path.isfile(os.path.join(TESSDATA_PATH, Languages[lang] + ".cube.lm")) or os.path.isfile(os.path.join(TESSDATA_PATH, Languages[lang] + ".unicharset")):
+ if os.path.isfile(os.path.join(TESSDATA_PATH, Languages[lang] + ".cube.lm")) or os.path.isfile(os.path.join(TESSDATA_PATH, Languages[lang] + ".traineddata")):
self.ListLanguages[lang] = Languages[lang]
self.cmbLang.append_text(lang)
@@ -1508,6 +1509,7 @@
self.LCfrench = "Francés"
self.LCitalian = "Italiano"
self.LCportuguese = "Portugués"
+ self.LCromanian = "Romanian"
#--- tips spanish
self.LCtooltips_tip = "Mostrar o esconder mensajes de ayuda"
self.LCselimg_tip = "Seleccionar las imágenes a convertir en texto"
@@ -1590,6 +1592,7 @@
self.LCfrench = "Français"
self.LCitalian = "Italien"
self.LCportuguese = "Portugais"
+ self.LCromanian = "Romanian"
#--- tips french
self.LCtooltips_tip = "Montrez ou masquer messages d'aide"
self.LCselimg_tip = "Sélectionner des fichiers d'images à changer en texte"
@@ -1672,6 +1675,7 @@
self.LCfrench = "Francese"
self.LCitalian = "Italiano"
self.LCportuguese = "Portoghese"
+ self.LCromanian = "Romanian"
#--- tips italian
self.LCtooltips_tip = "Mostrare o nascondere i messaggi di aiuto"
self.LCselimg_tip = "Selezionare le immagini da convertire in testo"
@@ -1757,6 +1761,7 @@
self.LCfrench = "French"
self.LCitalian = "Italian"
self.LCportuguese = "Portuguese"
+ self.LCromanian = "Romanian"
#--- tips english
self.LCtooltips_tip = "Show or hide tooltips"
self.LCselimg_tip = "Selecting images to convert to text"
@@ -1922,7 +1927,7 @@
#figure out where '/usr/share' is relative to the direcotry where we are located
SHARE_PATH = os.path.join(os.path.split(os.path.abspath(sys.argv[0]))[0], "..", "share")
TESSDATA_PATH = "."
- for path in ("/usr/share/tessdata",
+ for path in ("/usr/share/tesseract/tessdata",
"/usr/local/share/tessdata",
os.path.join(SHARE_PATH, "tesseract-ocr", "tessdata")):
if os.path.isdir(path):
Create romanian.patch in tesseract-gui-2.8/bin directory and apply.
patch -p0 < romanian.patch
Then install tesseract gui from tesseract-gui-2.8 with
make install
Romanian language should be available in the language selector from tesseract gui.
CORETECH 