dear Sven de Marothy,
You are doing great work with your camera I can understand from your report *Digitizing with a camera - some results.* ** I can confirm *all your findings* in your report regarding using digital camera for OCR.
Some years ago I used a Canon S40 4Megapixel camera and did som OCR of single book pages with an old version of FineReader.
I photographed maybe 50 or 100 big books, mostly handheld (camera and book) with day light from window. But very few of them I OCRed.
My experience was that the OCR result was as good as with a scanner, but if I wanted to do OCR also I had to spend more time photographing each book then if I just stored the books as pictures, so therefore I saved time by just going quickly through most of the books.
Now I have a Canon A620 7Megapixel and the results are just a little better. I also travel with a super small tripod and a big screw for attachement to tables (I put the book on the floor between the window and table).
I do a 60-page book in 5-10 minutes with a double page on each photo and about double time with a single page on each photo. The time also depends on how soft the book binding is and how good my working position is under the window.
I think the biggest problems with FineReader is that it does not accept scans that are not flat, and that it is sensitive to shadows. There is also a need for an expanded internal dictionary in Finereader. *An expanded dictionary that Finereader uses in the recognition phase (not later at spelling check). *Do you know if Finereader has this possibility?
I also need a Finereader that has built in recognition of diacritacal marks like you know åöä in Swedish. I do Indian language books.
mvh Mats Eklöf Huskvarna
2007/2/9, runeberg-request@lists.lysator.liu.se < runeberg-request@lists.lysator.liu.se>:
-
Mats Eklof