On 2012-06-07 13:37, Andrea Zanni wrote:
Well, that would be awesome, Gerard. OCR could be a critic feature of Wikisource (for example), the best softwares are all commercial (eg. ABBYY Finereader) and they cost a LOT. It would be very useful to understand what IMPACT does and where it is going. Moreover, Wikisource can be a testbed directly for OCR softwares, (because we have human proofreaders :-), and I think this is waht Remì was thinking :-)
It's great if Gerard can visit the meeting. I don't agree that the normal, commercial version of Finereader costs "a lot". If it did, free OCR software would have a fair chance, but Finereader is actually affordable, € 130 for a single person license. Most digital cameras cost more than that.
What IMPACT is doing, however, is to improve Finereader for blackletter/Fraktur, an area where both the standard Finereader and free OCR software do a poor job.
Wikisource can indeed offer the strength of manual, volunteer proofreaders in many different languages.