Anybody know any OCR that is actually any good at catching tables/flowcharts that kind of thing?

One that can handle Japanese as well would be nice, but probably asking too much

Thanks,

Ellen

The best OCR software is ABBYY FineReader. It supports 177 recognition languages. Unfortunately Japanese is not among them.

NiK

Thanks!

I downloaded the trial to see if it would catch the English parts of the bibliography and the numeric tables in the document I have been working on, but it didn't perform very well.

The quality of the image wasn't great though.

I found a free trial for Kanji OCR download which is for Japanese. I haven't tried it on any Japanese yet, but I will post what happens.

It won't read from PDF though.

Here is the link for anyone interested:

http://kanjikit.net/kocr/kocrdownload.htm

 I know this is an older post, but I got to it from Google and figured my input might be helpful to other users looking for something similar.

I haven't tried it with tables yet, but Adobe Acrobate Pro Extended [ and possibly the regular Adobe Acrobat version as well] has an OCR feature built into it that works surprisingly well. [ At least it has so far. ]  It recognizes both horizontal and vertical Japanese text. And because it's adobe, it works from PDF files. [ Although you can convert other formats into PDF pretty easily. ]

Hi Heather,

the problem with Acrobat is that it will only recognize documents with text layers (i.e. PDF exports, not scanned documents). Another problem: if you OCR a document in Acrobat and copy-paste the contents into, say, MS Word, each line will end with a paragraph break. Not very useful for CAT. Such documents need heavy pre-processing before you can start translating.

A more elegant, but still far from ideal, solution is using an OCR software, such as ABBYY.