by

Grabbing Text from Images with Project Naptha

I’ve lost a lot of hours to copying text from scans and photos of documents I’ve found on the web or taken for my own records. Some websites (particularly older websites that relied too much on flash or image files for their design) are particularly annoying to cite because nothing can be copied and pasted. While optical character recognition (OCR) software has come a long way, most of what I need isn’t worth the investment or the time of using specialized software. However, this might finally be getting a bit easier now that I’m using the new Project Naptha Chrome browser extension, which integrates a relatively smooth solution to the problem of text recognition into the browser.

Project Naptha (recently released by Kevin Kwok) looks through any image you view on the web for recognizable characters and converts them into selectable text, which can then be copied and pasted, translated, or to some extent edited. It has several limitations at this stage: I had to fix several errors in text I copied from scanned books, it will often pick up odd characters or artifacts when grabbing from something like a webcomic, and it doesn’t support text that is rotated too far off the horizontal plane. However, it’s so much faster than my previous process that fixing some errors is minimal clean-up–and it does a great job with grabbing text from memes and screenshots of old forums, which is worth it for studying digital culture alone. You can test a limited live demo on the website or download the free extension to try it out.

This project is another step towards centering a workflow right in the browser. Chrome is my favorite browser for its support of HTML5, and likewise a browser of choice for several of us at ProfHacker. If you’re looking to switch to Chrome, Amy recommended several extensions, including tools for Evernote and Zotero; Brian reviewed the clutter-reducing OneTab; and George recommended the accessibility extension Stylebot. Many of these extensions can keep a lot of everyday tasks right in the browser, which is of course also the vision behind Google’s Chromebook. (If you are interested in Chromebooks, Jason took a look at Samsung’s and Amy tried out ChrUbuntu and desktop Chrome Apps.)

Have you tried Project Naptha? What are your favorite tools for text recognition in image files? Share your experiences in the comments!

[Screenshot from testing Project Naptha with screenshots of Lewis Carroll's manuscripts from the British Library archive]

Return to Top