OCRFeeder (FOSDEM 2010)

static void_f_do_barnacle_install_properties(GObjectClass

*gobject_class){

GParamSpec *pspec;

/* Party code attribute */ pspec = g_param_spec_uint64

(F_DO_BARNACLE_CODE, "Barnacle code.", "Barnacle code",

0, G_MAXUINT64,

G_MAXUINT64 /* default value */,

G_PARAM_READABLE | G_PARAM_WRITABLE |

G_PARAM_PRIVATE);

g_object_class_install_property (gobject_class,

F_DO_BARNACLE_PROP_CODE,

Joaquim [email protected]

OCRFeeder

Documents conversion on GNOME

FOSDEM 2010

Joaquim Rocha (Igalia) · OCRFeeder · FOSDEM 2010

What is it?

Document Analysis and Optical Character Recognition

for GNOME


Why is it?

Paper has a number of problems

No applications for GNU/Linux to do a fair job


Paper problems:Security

CC Photo by: http://www.flickr.com/photos/badwsky/

http://www.flickr.com/photos/badwsky/


Paper problems:Preservation

CC Photo by: http://www.flickr.com/photos/98469445@N00/

http://www.flickr.com/photos/98469445@N00/


Paper problems:Data processing

CC Photo by: http://www.flickr.com/photos/hugovk/

http://www.flickr.com/photos/hugovk/


Paper problems:Ecology

CC Photo by: http://www.flickr.com/photos/pranavsingh/

http://www.flickr.com/photos/pranavsingh/


No fair conversion apps for GNU/Linux

apart from OCR engines, but...


OCR != Document Conversion

(it only deals with chars)(does not consider the layout)(does not distinguish contents)


what you want is

Document Analysis and Recognition

(conversion of documents to an electronic format)

(first projects in the 80s)


Where are were we at?

* Some closed solutions* Only for proprietary systems

* Various prices* still... arguable results


How?


How


Base concept:

1. Clip the contents2. Classify them

2.1. They are graphics → Paste on document

2.2. They are text → Calculate letter size; paste on document


So many layouts...

CC Photo by: http://www.flickr.com/photos/uber-tuber/

http://www.flickr.com/photos/uber-tuber/


Layouts vary with the type of document

What works on detecting one, won't work on others


OCRFeeder focus on contents, not on layouts!


Key concept:

If a document image can be divided in windows of 1 (content)

or 0 (not content), then it is possible to group all the

1s and outline the contents


Sliding Window Algorithm:

1. A NxN pixel window runs through the document top to bottom, left to

right



2. For every iteration, if there's a pixel inside the window which contrasts with the background,

then the window gets a 1,otherwise it gets a 0




It does not check all the pixels so there is a better performance



3. After all windows have a value assigned,

the ones with the value 1 are grouped



4. Every time a set of 1s is grouped, each window is reassigned the value 0

(these are called blocks)



5. When all windows have the value 0, the algorithm

reached the end


Block structure:


Joining Blocks:

Blocks are check with each other and joined when appropriate

When no blocks can be joined the analysis part is finished


Recognition:

System-wide OCR engines are used

Engines are configured from the GUI or XML files


Engine configuration:

<?xml version="1.0" encoding="UTF-8"?><engine>

<name>Tesseract</name><image_format>TIFF</image_format><engine_path>/usr/bin/tesseract</engine_path><arguments>$IMAGE $FILE; cat $FILE.txt;

rm $FILE.txt</arguments></engine>


Classification:

It is graphics if:

* Text is empty

* More than 50% of the chars arefailure chars, punctuation or spaces


Font Size Detection:

“Measures” in pixels the size of each text line

Although it results in different sizes...



The value equal or greater than the average is chosen

(results in values equal or close to the original font size)



The font size is calculated in inches using the resolution (DPI)

(if there's no resultion info, assume 300 DPI)

The value is then divided by the DTP (DeskTop Publishing point): 72


Exportation to ODT:

Uses ODFPy

(abstracts ODF creation)(just above XML)


User interaction:

Users can edit everythingand review the algorithm's results

So, UI can work in attended and unattended ways

CLI only works in an unattended mode


Nuance Omnipage test

ABBY Finereader test


Omnipage'sresults

Finereader'sresults


Demo time!


Other features:

* PDF importation* Unpaper preprocessor

* Font style edition* Exportation to HTML

* Project saving/loading


Future:

* Integrate Ocropus as an alternative analysis backend

* More exportation formats: HOCR, txt, PDF

* Improved a11y* Better integration with GNOME

and other GNOME apps


GNOME:

Development moved to GNOME's infrastructure since last month


Webpage:http://live.gnome.org/OCRFeeder

git:http://git.gnome.org/ocrfeeder

Bugzilla:coming soon...

http://live.gnome.org/OCRFeeder

http://git.gnome.org/ocrfeeder


Thank you!

Technology

OCRFeeder (FOSDEM 2010)