Apply metadata to files based on OCR of their content

Question

Essentially I'm looking for a tool (Windows/Mac) that will allow me the following workflow:

1. Scan ALL my documents into a folder (200-300 scanned images)
2. run the tool that will go through all the files and run OCR on them
3. based on the OCR, meta-data is applied onto each file.
4. I then read the meta-data and accordingly categorize files through a batch process.

While there are quite a few suggestions in SU & SE for doing plain OCR on files, I wasn't able to find a solution that essentially allows me to do programmatic like stuff based on the OCR-ed data from the documents.

The document template is standard so we know what kind of file to expect. We just want to scan the whole bunch and then run a backend process that neatly categorizes/uploads into respective folders. Having it OCR'ed gives me option to search within the file, while i open it in a program like Acrobat reader/Preview. But I want to run this categorization logic from a batch/shell/apple script. Stuff like prefixing the document number to filename etc.

Answer 1

What you describes seems to be a unique usage scenario. I don't beleive that there are solutions that work the way you describe out of the box. If this is a non-commercial/study project you don't have a much of a choise, espessialy for windows. Just google for tesseract an look what you can do.

If you have some budget, you have several choises. It's not very clear whether you need ready-to-go solution or a SDK, i assume you need the first one, so why don't you have a look at recognition server, it's a server-based OCR solution designed for mid- to high-volume document processing across large departments and enterprises. It can be either deployed as a standalone program or integrated with a third-party system such as DMS, RMS, and electronic archiving system.

I work @ ABBYY and can provide you more info if necessary. Hope it helps!