Training Tesseract - Revision history

Firespeaker: /* Fontproperties file */ realigned code blocks

2015-12-30T21:50:03Z

Fontproperties file: realigned code blocks

Vigneshv: /* Clustering */

2015-12-30T21:47:48Z

Clustering

Vigneshv: /* Fontproperties file */

2015-12-30T21:45:48Z

Fontproperties file

Vigneshv: /* Final steps */

2015-12-30T21:43:48Z

Final steps

Firespeaker: /* Training */

2015-12-30T21:42:22Z

Training

Firespeaker at 21:42, 30 December 2015

2015-12-30T21:42:00Z

Vigneshv: /* Clustering */

2015-12-30T21:41:57Z

Clustering

Vigneshv: /* Creating Training Images */

2015-12-30T21:38:54Z

Creating Training Images

Vigneshv: /* Final steps */

2015-12-30T21:38:32Z

Final steps

Vigneshv: /* Fontproperties file */

2015-12-30T21:38:05Z

Fontproperties file

@@ Line 44: / Line 44: @@
 Run the following 3 commands:
-    $ shapeclustering -F font_properties -U unicharset [lang].*.exp0.tr  # only for indic languages
+ $ shapeclustering -F font_properties -U unicharset [lang].*.exp0.tr  # only for indic languages
-    $ mftraining -F font_properties -U unicharset -O [lang].unicharset [lang].*.exp0.tr
+ $ mftraining -F font_properties -U unicharset -O [lang].unicharset [lang].*.exp0.tr
-    $ cntraining [lang].*.exp0.tr
+ $ cntraining [lang].*.exp0.tr
 Rename the files <code> normproto, pffmtable, inttemp </code> to be prefixed with <code> <lang_code>. </code>
@@ Line 53: / Line 53: @@
 For Tuvan:
-    $ mftraining -F font_properties -U unicharset -O lang.unicharset tyv.*.exp0.tr
+ $ mftraining -F font_properties -U unicharset -O lang.unicharset tyv.*.exp0.tr
-    $ cntraining tyv.*.exp0.tr
+ $ cntraining tyv.*.exp0.tr
-    $ mv normproto tyv.normproto
+ $ mv normproto tyv.normproto
-    $ mv pffmtable tyv.pffmtable
+ $ mv pffmtable tyv.pffmtable
-    $ mv inttemp tyv.inttemp
+ $ mv inttemp tyv.inttemp
 === DAWG files ===
 It is recommended that you have a list of word bigrams (line separated) and a wordlist (also line separated, but not necessarily complete). Run:
-    $ wordlist2dawg wordlist [lang].word-dawg lang.unicharset
+ $ wordlist2dawg wordlist [lang].word-dawg lang.unicharset
-    $ wordlist2dawg bigram_list [lang].bigram-dawg lang.unicharset
+ $ wordlist2dawg bigram_list [lang].bigram-dawg lang.unicharset
 For Tuvan:
-    $ wordlist2dawg wordlist tyv.word-dawg tyv.unicharset
+ $ wordlist2dawg wordlist tyv.word-dawg tyv.unicharset
-    $ wordlist2dawg bigram_list tyv.bigram-dawg tyv.unicharset
+ $ wordlist2dawg bigram_list tyv.bigram-dawg tyv.unicharset
 === Final steps ===
 To get the final <code>.traineddata</code> file, run:
-    $ combine_tessdata lang_code.
+ $ combine_tessdata lang_code.
 For Tuvan:
-    $ combine_tessdata tyv.
+ $ combine_tessdata tyv.
 The output will be <code> tyv.traineddata </code>

@@ Line 55: / Line 55: @@
     $ mftraining -F font_properties -U unicharset -O lang.unicharset tyv.*.exp0.tr
     $ cntraining tyv.*.exp0.tr
+    $ mv normproto tyv.normproto
+    $ mv pffmtable tyv.pffmtable
-And prefix <code> normproto, pffmtable, inttemp </code> with <code> tyv. </code>
+    $ mv inttemp tyv.inttemp
 === DAWG files ===

@@ Line 35: / Line 35: @@
 == Fontproperties file ==
-You must specify a font_properties file, with each line a font in the following format:
+You must specify a file named <code> font_properties </code>, with each line a font in the following format:
 <code> <fontname> <italic> <bold> <fixed> <serif> <fraktur> </code>, where you fill in each property with a 1 or 0 depending on whether the property exists. For example for Times new roman italic, a serif font:

← Older revision		Revision as of 21:43, 30 December 2015
Line 75:		Line 75:
	$ combine_tessdata tyv.		$ combine_tessdata tyv.

	The output will be <code> $ tyv.traineddata </code>		The output will be <code> tyv.traineddata </code>

@@ Line 17: / Line 17: @@
 The first step in training is generating tr files from the images you created. Do this by running:
-    $ tesseract [lang_code].[fontname].exp0.tif [lang_code].[fontname].exp0 box.train
+ $ tesseract [lang_code].[fontname].exp0.tif [lang_code].[fontname].exp0 box.train
 You will have to run this command for each font.
 For Tuvan and Times New roman:
-    $ tesseract tyv.TimesNewRoman.exp0.tif tyv.TimesNewRoman.exp0 box.train
+ $ tesseract tyv.TimesNewRoman.exp0.tif tyv.TimesNewRoman.exp0 box.train
 === Character set ===
 To get the charset, run:
-    $ unicharset_extractor [lang].*.exp0.box
+ $ unicharset_extractor [lang].*.exp0.box
 This gets all the box files, so you can run that command verbatim for all fonts.
 For Tuvan:
-    $ unicharset_extractor tyv.*.exp0.box
+ $ unicharset_extractor tyv.*.exp0.box
 == Fontproperties file ==

@@ Line 5: / Line 5: @@
 Tesseract has an option to generate images from text doing training. To do this, run:
-    $ text2image --text=training_text.txt --outputbase=[lang_code].[fontname].exp0 --font='Font Name' --fonts_dir=/path/to/your/fonts
+ $ text2image --text=training_text.txt --outputbase=[lang_code].[fontname].exp0 --font='Font Name' --fonts_dir=/path/to/your/fonts
 On Ubuntu, fonts are usually at <code>/usr/share/fonts </code>, but this path is platform specific. If you are training on multiple fonts, you will have to run this command multiple times. For the purposes of text2image, italics are considered a different font (you will have to run it once for Times, and once for Times Italic, for example)
@@ Line 11: / Line 11: @@
 For example, if you are training for Tuvan with Times New Roman:
-    $ text2image --text=training_text.txt --outputbase=tyv.TimesNewRoman.exp0 --font='Times New Roman' --fonts_dir=/usr/share/fonts
+ $ text2image --text=training_text.txt --outputbase=tyv.TimesNewRoman.exp0 --font='Times New Roman' --fonts_dir=/usr/share/fonts
 == Training ==