OSRA: Optical Structure Recognition

1. Description
2. Dependencies
3. Other acknowledgements
4. Compilation
5. Usage
6. License
7. Download
8. Web Interface
9. Author

Description:
OSRA is a utility designed to convert graphical representations of chemical structures, as they appear in journal articles, patent documents, textbooks, trade magazines etc., into SMILES (Simplified Molecular Input Line Entry Specification - see http://en.wikipedia.org/wiki/SMILES) or SD file - a computer recognizable molecular structure format. OSRA can read a document in any of the over 90 graphical formats parseable by ImageMagick - including GIF, JPEG, PNG, TIFF, PDF, PS etc., and generate the SMILES or SDF representation of the molecular structure images encountered within that document.
Note that any software designed for optical recognition is unlikely to be perfect, and the output produced might, and probably will, contain errors, so a curation by a human knowledgeable in chemical structures is highly recommended.

News:

Dependencies:
OSRA needs the following Open Source libraries installed:
Other acknowledgements:
OSRA also makes use of the following software (you do not need to install it separately, it's included in the distribution):
Compilation:
Unpack downloaded source code for the OCRAD package. Do not compile or install it - OSRA will automatically patch it and compile the object files it needs. Compile and/or install all the other necessary dependencies. Unpack OSRA package. Edit the included Makefile to make sure you have the correct locations for potrace, gocr, openbabel, and tclap. Check that GraphicsMagick++-config location (it's a script that comes from GraphicsMagick installation). You might have to set LD_LIBRARY_PATH to /usr/local/bin or wherever you have installed OpenBabel. Set ARCH variable to one of the following: unix - for linux,unix,osx; win32 - for building on Windows MinGW environment. Running make should then generate the executable - osra.
More detailed instructions for compiling OSRA for OS X and Windows platforms are provided in the README file.
Usage:
OSRA can process the following types of images: Some common abbreviations, hetero atoms, fused and merged atomic labels, hash and wedge bonds, and bridge bonds are currently recognized. Formal charges, isotopes and some element symbols, i.e. iodine ("I" -- looks too much like a straight line = single bond), are not.

Command-line options:
./osra --help
will give you a list of available options with short descriptions.

Most common use: ./osra [-r <resolution>] <filename>
Resolution in dpi, default is 300 (unless it's a PS or PDF file as mentioned above), filename is the name of your image file (or PS/PDF document).

Other options:
-t, --threshold: Gray level threshold, default is 0.2 for black-and-white images,
-n, --negate: Inverts colors (for white on black images),
-o, --output: Sets a prefix for writing recognized images to files - i.e. "-o tmp" will create files tmp0.png, tmp1.png... for each of the structures,
-s, --size: Resize images on output - can be useful for running OSRA as a backend for a webservice. Example: "-s 300x400".
-g, --guess: Prints out resolution guess when you chose to have automatic resolution estimate.
-p, --print: Prints out the value of confidence function estimate.
-f, --format: Output format (either smi for SMILES or sdf for SD file format)
-d, --debug: Print out debug information on spelling corrections. First column - output from the OCR engine, second - result of spelling correction, last - SMILES from the superatom dictionary, if any.
-a configfile, --superatom configfile: Superatom label map to SMILES (superatom.txt by default)
-l configfile, --spelling configfile: Spelling correction dictionary (spelling.txt by default)
-e, --page: Show page number for structures from multi-page PDF and PostScript documents
-R, --rotate: Rotate image clockwise by the number of degrees, i.e. -R 90

License:
This program is free software; the part of the software that was written at the National Cancer Institute is in the public domain. This does not preclude, however, that components such as specific libraries used in the software may be covered by specific licenses, including but not limited to the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version; which may impose specific terms for redistribution or modification.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA. See also http://www.gnu.org/.

See the file COPYING for details.

Download:
OSRA is Free and Open Source Software. You are welcome to download and use it, provided that you understand the terms described above. Participation in the development is highly encouraged!
Download OSRA
We also welcome your feedback - send us your comments, suggestions, criticism, or praise to the contact email address below.
Web Interface:
To demonstrate the capabilities (and limitations) of OSRA we have created the following web interface:
OSRA Web Interface
Try this sample image from the US Patent Office website first: patent.gif. Use a resolution of 300 dpi.
Author:
Igor Filippov, igorf(AT)helix.nih.gov
2007-2008, SAIC-Frederick, NCI-Frederick, NIH, DHHS, Frederick, MD
SourceForge.net Logo