Copyright©2008 Antony Williams
I have posted a number of blogs previously about chemistry document markup and our efforts in this area (1,2,3) then last week announced ChemMantis, our Chemistry Document Markup alpha-release. In the original presentation I gave on our document markup system at the ACS in Philly (online here) I talked about he possibility of integrating optical structure recognition tools. These tools are software packages/components that convert structure drawings to connection tables (4,5). I have discussed these previously on this blog in terms of my work with CLiDE (6,7) and with OSRA (8).
OSRA is an open source package for Optical Structure Recognition developed by Igor Filippov at the National Cancer Institute. My early experience with OSRA wasn’t all positive (8) but since it is Open Source we have integrated the latest software to ChemMantis and we have been testing it out. There are instances where the software works perfectly and the structure generated from the image is perfect and there are examples where the conversion fails. Examples of both are shown below. The top image shows an incorrectly converted image and the bottom one a correctly converted image. At present it is clear that such conversions should be inspected by the user and edited if necessary. OSRA certainly offers an opportunity to shortcut the drawing of chemical structures.Stumble it!