Automatic file name with OCR when scanning

[fj.leon]

Hello, we are new at openkm and want to use it in our company. We want the scanner tool inside openkm to NOT request the file name before scanning, rather the document should be scanned and uploaded and with OCR it should analize a specific region in the document and use the detected text as the file name.

My boss asked for this and i think it's a long shot, i don't expect it to work but i must ask before giving him the answer.
The reason for this is to avoid user error when entering the file name, since we must rely on it for the company procedures.

[jllort]

Do not exists any Free OCR engine which you can use to scan zones, basically it's the problem. The only aproximation could be that in all document could be some tags, for example filename_start doc_name filename_end and that could be parsed before ocr engine, but forget in your mind any free solution with zone scanning.

You could try with this open source http://code.google.com/p/ocropus/ based on tesseract , but I think do not working with zones. If you found some one tell us well be pleased shared this kind of information with community.