Taking the pain out of translating text in images
Written by ForeignExchange Translations on Thursday, November 12, 2009
When translating CMC documentation, SOPs, or regulatory submissions, we regularly run into graphics that contain text to be translated. Translating text contained in images can be a slow and frustrating task.
The main problem? Source files that are not available to translators. So we're often left with manually extracting text and then recreating or editing images.
One strategy for making images easily translated would be for drug and device companies to more broadly support the open standard SVG format. Text in an SVG image can be changed without changing the rest of the image because an SVG file is itself a text file.
One of the main advantages of SVG is that the image file itself is the source file. However, SVG as a format is not widely supported by many applications. Therefore, it is often necessary to export SVG to a supported format to use it on a web page - which leaves us with the same problem of keeping and then making available to translators the source SVG files.
Despite the fact there is no "easy fix", there are some good tools out there to help with localizing. One such tool is Bjorn Austraat's Globalization Image Assistant (GIA).
The tool can speed up the time needed to the task of analyzing, preparing, and QAing graphic images. Two items that would make GIA even more useful would be if it included the actual graphic in the exported reports and if it had a tie-in to OCR. GIA is definitely worth the investment!
However, to address our specific needs, the ForeignExchange team ended up developing a macro that extract graphics into a Word table with the file path, graphic image, and 2 columns for entry of source/target text. To populate the source text, we use the following process:
- Create a DOS directory listing of all graphics.
- Turn the listing into an HTML page with links to all the graphics.
- Print the HTML page to the Microsoft Office Document Image Writer print driver, which then opens in the MS application that allows you to OCR the text.
- Select Tools -> Send Text to Word.
Dealing with scanned images and text is a fact of life in the medical translation space. However, GIA or homegrown solutions can help you take some of the frustration out of the process!
UPDATE: A previous version of this article mentioned a $99 price for GIA. It seems that was an introductory offer.
Contact ForeignExchange to find out more about our specialized medical translation services for CMC documentation, SOPs, and regulatory submissions.
Also, I just wanted to mention just because a file has an .svg extension does NOT mean the text will be editable. There is a good chance it will be, but something to be aware of. There is not a single graphic extension that can guarantee without double checking that the text will be editable in the source.
Thats a lot of money.
J Nelson Rn BSN